No.

17 - June 1979

Shells Around Southern Novae

H. W. Duerbeck and W. C. Seitter

Spencer Jones published in 1931 a bulky volume of

Although less spectacular than their big brothers, spectroscopic and visual observations made at the Cape
the supernovae, the novae are by no means less Observatory. Another southern observer, J. Hartmann in
Buenos Aires, observed the nova spectroscopically and
in te res ting. They are also much more frequent
wrote the most concise astronomical paper ever publish-
and several are known in the southern sky. After ed, a telegram senJ to the Astronomische Nachrichten:
the initial explosion, a shell expands around the "Nova problem solved; star blows up, bursts." And indeed,
nova and may become visible after a while. Drs. when double-star observers examined the postnova two
Hilmar Duerbeck and Waltraut Seitter from the
Hoher-List Observatory, near Bonn, FRG, recently
observed three southern novae. The excellent
resolution of the 3.6 m photos makes it possible to
see details in the very faint nova shells that have
never been perceived before.

The southern sky comprises one of the most fanciful

supernova remnants-the extended spider web of the
Gum nebula. It harbours also some less spectacular, tiny,
astronomically shortlived phenomena: the remnants of
near nova explosions. They can be observed for only a few
decades after outburst, before they thin out and merge into
the interstellar medium. Due to their small size and low
surface brightness, they require large telescopes, such as
have recently become available in the southern hemi-
sphere. Fortunately, some observing time was granted to
us before the above-mentioned disappearances!

RR Pictoris
Two brilliant novae shone in the southern sky in this
century. The first one, RR Pie, was discovered on May 25,
1925, and reached its peak magnitude of 1~2 on June 9. It Fig. 1: The nebular shell around T Pyx (1966). From a prime foeus
remained visible to the unaided eye for about a year. 098 plate behind a RG 630 filter, exposure 70 min.

Fig. 2: The nebular shell around RR Pie (1925). From a3.6 m prime
foeus 098 plate behind a RG 630 filter, exposure 90 min.
years later, it was embedded in a nebula, with two distinct
concentrations moving in opposite directions. We quote:
"1928.284. Nebula slightly elongated in 250 o ±. Does not
merely look double; looks granular and separated by
patient watching. AC 253°±, AB 124°±, AD353°±.
x 810. In my opinion, A, B, C certain, D somewhat
doubtful ... " (Spencer Jones 1931).
Have a look at the appearance of RR Pic on a 90 rn red
exposure, taken at the prime focus of the ESO 3.6 m
telescope (fig. 2). The postnova is su rrounded by a structu-
red nebulosity. A ri ng-I ike feature encompasses the central
star, two pairs of condensations seem to have been ejected
with higher (tangential) velocities, at right angles to the
ring and opposite to each other. They can easily be
identified as fragments Band C of the early double-star
observers. Ais, of course, the central star, and D part of the
ring-like structure. The computer-enhanced, blue image is
shown in figure 3.
RR Pic resembles the well-known remnant of Nova DQ
Her, which also displays an "equatorial ring" and "polar
blobs", but it has a more complicated structure: the polar
blobs are double or even tri pie, and there are two
equatorial rings, inclined to each other, and each nearly
perpendicular to an axis joining prominent polar conden-
sations. The similarities between DQ Her and RR Pic are
striking: both are slow novae, both are short-period
binaries (Walker 1954, Vogt 1975), both nebulae show the
same basic geometry.
An important, and in principle straightforward, use of
a nova remnant photograph is the determination of the
2
(
Fig.3: The nebular shell around RR Pie (1925). Fromaprime foeus
111 a-J plate behind a BG 385 filter, exposure 60 min. Plot prepared
from POS density seans.
nebular parallax: with the known expansion velocity of the
shell in km s " as observed during outburst, and the
dimensions of the shell in arc seconds at some later time,
the distance can be determined without any further
assumptions. But alas! RR Pic showed many radial velocity
systems in the course of its evolution, ranging from 40 to
1,600 km s " and the shell has many diameters: that of the
ring, that determined by the blobs, or merely, the projec-
tion of the blobs in the sphere, since the open ring
structure indicates that the polar blobs are not ejected
perpendicular to the line of sight. A preliminary analysis
leads to a distance of 400 pc. This is a slight revision of the
previously assumed value.
CP Puppis
The nova with the fastest development, except V 1500 Cyg
1975, is CP Pup of 1942. Observed expansion velocities
were of the order of 1,200 km s '. High-resolution spectra
obtained in the later nebular stage by Sanford (1945)
showed Hß and [0111] 4364 broken up into more than
a dozen emission components.
Zwicky, in 1955, obtained a practically featureless photo-
graph of the shell with the Palomar 5 m reflector. When he
published the photograph in 1962, he wrote: " ... it would
merit greater attention than it has been accorded hitherto
and it is to be recommended for more observations
particularly to observers in the southern hemisphere ... "
Here it is (fig. 4): A fringed halo on the blue plate, a chain of
black pearls on the red plate: a late confi rmation of the
early fragmented emission line profile? Again it must be
said that a straightforward application of the nebular
expansion parallax is not possible, but if it is assumed that
the major portion of the material was ejected in a slightly
inclined ring, data from the spectroscopic study and the
direct photograph can be reconciled, The derived distance
of 1,500 pc is in excellent agreement with an earlier
determination, based on galactic rotation and interstellar
lines, and leads to a very high peak brightness of M =
-11~5. CP Pup was very likely the most luminous nova
observed until now.
T Pyxidis
The third remnant, the nebulosity around the recurrent
nova T Pyx, was in same respects a surprise, Recent
photographs of the brightest recurrent nova, T CrB, had
revealed only very weak nebular whisps (Williams 1977).
The very strang nebulosity around T Pyx is thus unusual.
Again it is not trivial to derive the nebular parallax, Radial
velocity observations are scarce for T Pyx and we find the
added problem of having to decide which outburst caused
the remnant. Was it produced in 1966, 1944, 1920, 1902, or
even 1890?
Fortunately, Catchpole (1969) provides us with the
knowledge of a radial velocity system observed in 1966 of
v = -900 km s " interpreted as being due to the principal
spectrum. With this, a distance of 600 pc is deduced,
corresponding to absolute magnitudes at maximum of Mv
= 2'~9 and at minimum of Mv = + 4':"4. Under the above
assumptions, the recurrent nova does not fit into the.
t 3-absolute magnitude at maximum-relation.
It is possible to check on the 1966 origin ofthe shell. With
the distance and the angular diameter known, the volume
enclosed by a spherical nova shell and its content of
interstellar matter can be deduced. Assuming that the
detection of a shell requi res its density to be 10-100 times
Fig, 4: The nebular shell around CP Pup (1942). From a prime
foeus 098 plate behind a RG 630 filter, exposure 60 min.
New ESO Slide Sets
Two new ESO slide sets will become available during the next
months.
The first of these consists of 20 5 x 5 cm colour slides
showing the ESO installations on La Silla. Buildings, tele-
scopes and views of the site are included. A full description in
several languages explains the slides.
The second set contains some of the best photographs that
have been obtained with the 3.6 m prime focus camera
(Gascoigne corrector). 20 black-and-white slides have been
selected from more than 1,000 photographs, Nebulae, gal-
axies, etc. Full details in accompanying text.
The price for one slide set is German Marks 18,- (or the
equivalent) for Europe, and US$ 10,- by surfacemail to all
other countries, or US$ 12.50 by airmail (to be paid in
advance),
Send your cheque or bank drall to:
European Southern Observatory
Schleissheimer Strasse 17
0-8046 Garehing bei München
(Commerzbank, München, Account No, 2102 002).
So me copies of the first ESO slide set with 20 photos from the
1 m Schmidt telescope are still available (same price as above),
the interstellar density, we find that a 1944 shell must
contain at least 1026_10 29 9 of matter and, with a suspected
higher expansion velocity of 1,700 km s ',even 1029_10 30 g,
while a shell ejected in 1966 has a lower mass limit of
1027 -10 26 g, The former values seem rather tao large,
especially in view of earlier determinations for other
recurrent novae which yield 1026 9 or less. It must be kept in
mind, however, that the above argument requires spherical
volumes. Thin shells or remnants with strang condensa-
ti ans combine smaller ejected masses with langer life-
times,
Perhaps the strongest argument in favou r of a 1966 shell
is the absence of strang remnants of earlier outbursts,
possible proof of a short lifetime, as weil as a more
spherical nature of recurrent nova shells.
While all arguments are weak, they provide us with
a comfortably complete set of data for a 1966 outbu rst.
With the implied fast development of the remnant it is
easily possible to verify or to reject our conclusion by
systematically following the future development of the
remnant of T Pyx,
The photographic investigations reported here are
clearly only the beginning of a closer study of nova
remnants and must be supplemented by spectroscopic
investigations which will permit a closer look atthe physics
of the nova shell and hopefully shed more light on the
natu re of the nova process.
We are pleased to thank the night assistant of the ESO
3.6 m telescape, Sr. Yagnam, for most efficient handling of
the telescape and the coffee machine, and Dipl. phys. H.J,
Becker (Bann) for his readiness to produce beautiful plots
from POS scans,
References
Catchpole, R,M" 1969, Mon. Not, 142,119,
Hartmann, J" 1925, Astr. Naehr. 226,63,
Spencer Jones, H., 1931, Ann. Cape Obs.10, part 9,
Sanford, R. F" 1945, Astrophys, J,102, 357,
Vogt, N. 1975, Astr, Astrophys.41, 15,
Walker, M. F., 1954, Publ. Astr, Soe. Paeifie 66,230.
Williams, R.E., 1977, Veröff, Sternw. Bamberg9, Nr.121, 242,
Zwicky, F" 1965, KI. Veröff. Sternw. Bamberg 4, Nr. 40, 169.
3
Mapping of Galaxies at High Radio Frequencies
R. Wielebinski

the Galaxy. The strongest of these SNR's should be easily

Professional astronomers are sometimes asked
the question of whether observations with radio
telescopes are better than the "old-fashioned"
optical observations? The answer is of course that
they are equally valid: to understand the objects in
the Universe, we must observe them over the
widest possible spectral range. For this purpose,
short-wavelength observations are now carried
out from spacecraft, and in the other end of the
spectrum we rely upon the ingenuity of the radio
astronomers with their giant antennas. In this
review, Dr. Richard Wielebinski ofthe Max Planck
Institute for Radioastronomy in Bonn, FRG, gives
examples of the important interaction between
optical and radio observations for the study of
nearby galaxies.
detected at a distance of even 4 Mpc as individual radio
sources. The supernova events are known to produce
pulsars, and they as weil release energy in the form of
relativistic particles. These electrons in turn produce
diffuse nonthermal emission which is linearly polarized.
The measurement of the linear polarization should enable
us fu rthermore to study the magnetic fields in the galaxies.
Sensitive measurements of the emission above the
plane, particularly in edge-on galaxies, allow us to study
the diffusion (or convection) of relativistic particles from
the sites of their formation into the magnetic fields of
a possible "halo". The study of the diffuse thermal
emission, known to exist in our galaxy from the absorption
of low frequency continuum emission, could be tackled
once a careful separation of the thermal/nonthermal
emission is made. To separate all these effects radio maps
at many frequencies are required. These radio results,
combined with various other observations, should enable
us to understand the energy balance of galaxies.

Mapping the Spectral Indices and Magnetic Fields

The detailed study of the distribution of the radio conti-
nuum emission of "normal" galaxies began only ten years
ago, when G. G. Pooley published a map ofthe Andromeda
nebula (M 31) made at 408 MHz. This map showed for the
first ti me details of the spiral structure at radio frequencies.
Until thattime no single-dish radio telescope had sufficient
angular resolution and no synthesis array the necessary
brightness sensitivity to be able to map nearby normal
galaxies. A typical angular resolution required for the
largest objects is a few minutes of arc, while tens of
seconds of arc resolution allow us to study numerous
smaller galaxies. A number of presently operating synthe-
sis arrays have this resolution, and have been used to map
galaxies, particu larly at lower radio frequencies. The
surface brightness of the radio continuum emission of
normal galaxies is low, typically a few degrees K at
408 MHz, but dropping rapidly to a few mK at 4800 MHz.
(The temperature spectral index ß is typically - 3.0 for
galaxies in this frequency range - T cx: VB.) For such
sensitive measurements the large collecting area of
a single dish, like the 100 m Effelsberg radio telescope, is
ideal. To map galaxies at frequencies above 5 GHz, where
weather effects seriously hamper observations, the
development of new techniques was necessary to allow
studies to be made in this important frequency range.
The 100 m radio telescope of the Max-Planck-Institut für
Radioastronomie has been used to map nearby galaxies at
a number of frequencies from 840 MHz to 23 GHz. At first
a I.. 11 cm (2.7 GHz) map of M 31 was made with r. m. s.
noise of 3 mK. Measurements with such sensitivity were
never made before and were only possible due to the
combination of the excellent telescope and the highly
stable, low-noise receiver. The data obtai ned for M 31 were
studied in detail by E. Berkhuijsen, particularly for correla-
tions between radio continuum and the various consti-
tuents like H I1 regions, OB associations, supernova rem-
nants, H I gas, blue light, etc. Similar investigations have
been made on the basis of I.. 11 cm and 1..6 cm maps of M 33.
In M 33 the emergence of the thermal emission as the
dominant constituent, even below 4.8 GHz, is evident.

Thermal and Nonthermal Emission

To understand the importance of mapping of galaxies at
high frequencies, a short summary of the emission pro-
cesses which produce the radio continuum in our galaxy
should be made. Along the galactic plane we have a narrow
band of discrete H 11 regions with thermal (flat) spectrum.
The brightest of these H 11 regions placed at the distance of
1 Mpc would be barely detectable as individual sources,
but the integrated effect should certainly be the dominant
emission at the highest radio frequencies. A somewhat
broader distribution of nonthermal supernova remnants Fig. 1: The 100 m radio lelescope al Effelsberg. 11 has been used 10
(with steep spectrum) is found along the galactic plane in map galaxies al frequencies as high as 23 GHz.

4

Fig. 2: M31-An overlayofradiocontoursontoa LickObservatory
photograph. 2695 MHz (;.. 11.1 cm), 4~ 7 beam (Berkhuijsen and
Wielebinski).
Studies of IC 342, M 81, M 51 and, more reeently, M 82 at
frequeneies as high as 23 GHz have aimed at an accurate
thermal/nonthermal separation. To map galaxies at these
high frequeneies, without being subjeet to base-level
variations due to perturbations from atmospherie thermal
emission, the teehnique of beam switehing was extended
from point sources to extended objeets by D. T. Emerson,
U. Klein and G. Haslam. It now appears possible to map
..
•
Fig. 3: NGC 253-An overlay of radio contaurs onto an ESO Schmidt plate. 8.7 GHz (;..3.4 cm). The radio emission shows an extended
halo surrounding the galaxy (Beck, Biermann, Emerson and Wielebinski).
galaxies with 1 mK r. m. s. noise and aresolution of 1.2 are
min at 10.6 GHz (A 2.8 em). In the near future maps with
- 30 are sec resolution at 32 GHz (A9.6 mm) should be
possible. Such maps, when eombined with similar resolu-
tion maps made, for example, with the Westerbork Synthe-
sis Radio Teleseope in the Netherlands at 61 0 or 1420 MHz,
or with the Cambridge 150 MHz teleseope in England, ean
give us exeellent speetral index distribution maps and
henee thermal/nonthermal ratios in a number of nearby
galaxies.
Maps of linear polarization of M31 and M33 have been
made by R. Beek at 2.7 GHz with 4.4 are min angular
resolution. These observations have shown that the non-
thermal emission is generated in magnetie fields whieh are
ordered on seales of kpe. The degree of polarization is
high, up to 40 % in some areas. As yet no maps at other
frequeneies have been made, but in the south-preeeding
arm of M 31 there seems to be very little Faraday rotation.
The well-aligned "E" veetors imply that there is a large-
scale magnetie field of some 5 I-LG along this spiral arm.
Observations now planned at other frequeneies should
give details of the magnetic fields and of the electron
densities in the spiral arms.
Galaxy Halos
Nonthermal emission from the Galaxy was the first radio-
astronomical observation made some 40 years ago. Stu-
dies of nonthermal emission in our galaxy has led to
' .
..
5
a long-standing controversy about the existence of an
electron halo. If cosmic rays are to be contained in the
Galaxy, a rather strong electron halo is expected. The
refinement of measuring techniques over the last 20 years
has led radio astronomers to conclude that any such large
scale component surrounding the galaxy must be weak.
Also the experimental evidence indicates that the spectral
index of the radio continuum emission at high distances
above the galactic plane is steeper than the spectral index
near the plane.
Observations of edge-on or nearly edge-on galaxies
ofter the best opportunity to study the halo phenomenon.
An analysis of 408 MHz observations led R. Wielebinski to
conclude that any halo around M 31 is weaker than that
surrounding the Galaxy. Further studies of the halo of M 31
using the 100 m telescope were made recently atthe "Iow"
frequency of 842 MHz by R. Gräve et al. Other edge-on
galaxies mapped were NGC 891 and NGC 4631. The only
high frequency halo so far found was at 8.6 GHz in
NGC 253 (R. Beck et al.). This observation implies a young
population of relativistic electrons and this may be due to
the high nuclear activity seen in NGC 253. The relation
between nuclear source and radio emission in the disk is
unclear at present. Studies of the nuclei of galaxies,
particularly with the highest angular resolution of the VLBI
technique, should tell us something about these relations
and hence about the energy production in galaxies.
Wh at Comes Next?
Future instrumental developments necessary in this field
of research are now becoming apparent. Single-dish maps
at the highest frequencies will be able to provide informa-
tion on the thermal emission distribution in nearby gala-
xies. At present, to map a larger galaxy at 10.6 GHz down to
the confusion level of the telescope would take
ö Crucis is Variable!
E. w. Eist
During arecent visit to La Silla, Dr. Eric W. EIst of
the Royal Observatory at Uccle, Belgium, discov-
ered that one of the stars in the Southern Cross is
variable. So are many other stars, but the present
case is particularly interesting because the maxi-
mum amplitude in the lightcurve is only 0':'0061
The discovery is a powerful demonstration of the
quality of the La Silla site and a tribute to the
Bochum 61 cm telescope and its photometer.
Although the bright, southern star ö Crucis (V = 2'!'8) has
been observed many times during the past, its variability
has remained undiscovered until now. Due to the high
precision of the Bochum 61 cm photometric system and
extremely good weather conditions, it was possible, during
my last stay at ESO in February 1979, to detect a short-peri-
odic light variation of Ö Crucis, with an amplitude of only
0'!'006!
6
- 1,000 hours. Developments which would speed up the
observing, such as use of multi-beam receiver systems, are
highly desirable. Aperture synthesis telescope maps at
lower frequencies require the filling of the missing
spacings. If this is not done, the extended structure is lost
and the resulting map unusable for detailed studies.
Combination of synthesis arrays and single-dish maps
would give data which could be used in detailed spectral
studies. The improvement in VLBI sensitivity by the use of
a broader bandwidth should enable detailed studies of
a larger number of nuclei of galaxies.
Studies of the radio continuum distribution in galaxies
require parallel information from all other astronomical
observation modes. For the investigation of the thermal
content Ha-data are required. To study the relation of radio
continuum to the density wave theory, high resolution H I
studies of the same galaxies are needed. The halo of our
galaxy can be investigated either in radio continuum or in
y-ray observations. Molecular line studies of normal
galaxies have so far been limited either to nuclei or to
extremely large H 11 regions. The advent of new techniques
in all fields of astronomy and their application to studies of
nearby galaxies will certainly bring us"'nearer to under-
standing the workings of these beautiful beings.
General References
Kruit, P. C. van der, Allen, R.J. "The Radio Conlinuum Morphology
of Spiral Galaxies", 1976, Ann. Rev. Astron. Astrophys. 14,417.
Berkhuijsen, E. M., Wielebinski, R. (Eds.) "Structure and Proper-
lies of Nearby Galaxies", lAU Symp. No. 77, 1978, D. Reidel
Publishing Co., Dordrecht.
Other References
Berkhuijsen, E. M. 1977, Astron. Astrophys. 57, 9.
Pooley, G. G. 1969, Monthly Notices Roy. Astron. Soc. 144, 101.
Wielebinski, R. 1976, Astron. Astrophys. 48,155.
Fig. 1: The Southern Cross above La Silla. Ö Crucis is indicated.
Photographed by ESO photographer B. Dumoulin in 1977. The
two bright stars below are a and ß Centauri.
Earlier Observations amplitude of the Iight variation of b Cru. Table 1 gives some
information about the observed minima. The deduced
In 1956 the radial velocities of seven bright southern B-type period is P = 3~62491 and the amplitude is ß V = 0'!'006.
stars (from the D. H. McNamara list) were systematically
examined at the Radcliffe Observatory, in an effort to Table 1: Minima of b Cru
detect the presence of short periods (Pagel, 1956, MNRAS.
116, 10). Positive results were obtained for ß Cru, "[' Lup and Date (UT) UT
a Lup, whereas no conclusion could be drawn for Ö Cru
because the lines in the spectrum were too diffuse to be 31-1-79 7" 43 m
measu red accu rately. 12-2-79 6" 16m
It is interesting to note, that photoelectric observations 13-2-79 7" 22 m
of a Lup at that time by A. B. Muller (Leiden Observatory 15-2-79 6" 34 m
Southern Station) did not reveal any variation in excess of
0~01. a Lup was therefore considered as apparently
constant, which was confirmed by earlier observations of
this star, carried out at the Cape Observatory in 1948 and Figure 2 shows the lightcurve as deduced from observa-
1950. tions on two nights. Overlapping points are not shown.
In connection with an investigation of the multiperiodici-
ty of ß Cru (Van Hoof, 1959, Z. f. Astrophys.• 47, 198),
~
Haffner observed this star during several nights at the 1n0
BOCHUM 61 (m Cru
• 11-2-79
Boyden Observatory. He used Ö Cru as the comparison star ·13-2-79

during the first two nights, but its brightness was not
checked for constancy.
From the third night on, Ö Cru was replaced bythe closer 2780
...
star 39 Cru. This change, Van Hoof states, was not .....
a favourable one, since night-to-night variations were
found in the average ßm (ß-39).
""0
In 1972, Shobbrook (MNRAS. 156, 5P) again observed
Ö Cru, together with ß Lup. TJ Lup, Ö Lup and E Cen. After
a few nights it became apparent that Ö Lup va ried by up to 00 os 10

0~005 and E Cen by up to 0~015. The other three stars
appeared constant to 0~003, and were not observed
further.
In 1973, b Cru appears once more in a list, this time
compiled by Percy (A&A. 30, 465) with the aim of doing
some statistics of undiscovered ß Cephei stars. However,
since Percy relied on the Shobbrook investigation, Ö Cru
was considered as "not variable".
Finally Jerzykiewicz and Sterken (1977, Acta Astronomi-
ca. 27, 365) put Ö Cru on their list, during a search for
ß Cephei stars. But for some reason, they indexed this star
with "NO". This symbol indicates stars which the authors
do not plan to observe in their present programme, either
because they are already well-known variables, or because
no suitabel comparison stars could be found for them.
Fig. 2: The lighlcurve ofb Crucis, as measured in February 1979.
So, by pure coincidence, Ö Cru was now observed once
more, and due to thefine photometer at the Bochum 61 cm
telescope, and the excellent weather conditions, it could
no longer hide its variability!
'" '" '"
This article is dedicated to Mrs. Hilde Fritsch, formerly in
charge of the ESO Guesthouse in Santiago and now
retired. I also thank Mr. De Kersgieter for drawing the
figure.

Photometrie Observations with the Boehum 61 em

Teleseope
In order to establish more firmly the TPA (timeshift-period-
amplitude) relation (Eist, 1978, Astrophys. J., 223, 959),
I planned to observe in February 1979 several Ö Cephei
stars and some well-known ß Cephei stars. The observa-
tions were done with the automated photometer which is
attached to the Bochum 61 cm telescope. Due to the
extremely favourable weather conditions, I soon ran out of
stars! Therefore, without previous planning, I selected six
early B-stars from the sky in an arbitrary way, and observed
them for several nights. During the day time, a preliminary
photometric reduction of the observations was made by
means of a pocket calculator.
It immediately became clear that all of the six stars had to
be considered as variable (Eist, 1979, Inf. Bull. Var. Stars,
1562), but since at that time I did not know of the previous
history of Ö Cru, I did not observe this star very extensively.
However, from the observations of four minima and one
maximum, it is still possible to estimate the period and the
List of Preprints
Published at ESO Scientific Group
March-May 1979
51. G. TENORIO-TAGLE, H. W. YORKE and P. BODENHEIMER:
The Gas Dynamics 01 H 11 Regions. 111. Submitted to Aslronomy
and Aslrophysics.
52. D. PELAT and D. ALLOIN: High-Resolution Profile 01 the [0 111]
Lines in NGC 1068. Submitted to Aslronomy and Aslrophy-
sics.
53. G. CONTOPOULOS: How lar do Bars extend? Submitted to
Aslronomy and Aslrophysics.
54. M. AZZOPARDI and J. BREYSACHER: More Wolf-Rayet Stars
in the Large Magellanic Cloud. Submitted to Astronomy and
Aslrophysics, Suppl. Series.
55. M. P. VERON and P. VERON: A Study 01 the 4C Catalogue of
Radio Sources between 20° and 40°. 11 The sampie. Submitted
to Aslronomy and Astrophysics Suppl.

7
Cluster Hunt in the Southern Milky Way
L. 0. Loden

inspector's eye. Already in the beginning we were con-

Most ot the known stellar clusters in the Milky Way
have been tound because they contain conspicu-
ous groupings ot relatively bright stars. Nobody
.doubts, however, that there are many other clus-
ters, in particular very loose ones, wh ich are not
known at present. But how to disco ver them and
to prove that they are real physical entities? In this
article, Dr. Lars Olot Loden trom the Stockholm
Observatory, Sweden, summarizes one aspect ot
a large investigation that has been underway tor
nearly two decades: tinding new clusters in areas
that are densely packed with stars.
In 1962 we began to work seriously with the material from
the large spectral survey of the Southern Milky Way that
was outlined by the late Bertil Lindblad a few years earlier.
This material was secured at the Boyden Observatory in
South Africa between 1958 and 1962. It consisted of
objective-prism plates obtained with the ADH Baker-
Schmidt telescope (widened spectra), direct-photo-
graphic plates in blue and yellow taken with the 25 cm
Metcalf astrograph, and photoelectric UBV photometry of
a large number of standard sequences distributed overthe
whole investigated part of the sky (obtained with the
reflecting telescope). The region covered included a galac-
tic belt between I = 235 0 and I = 100 with an approximate
width of r5.
The intention was to select "interesting" stars from the
objective-prism plates by means of visual inspection.
These stars shou Id then be subject to subsequent, detailed
investigation, maybe also statistical studieswith respectto
spatial distribution, etc. A problem was that from the
beginning of the project it was not clear how to define the
concept "interesting stars"! Rather soon, however, it
turned out that the quality of the objective-prism ex po-
sures was relatively uneven, ranging from excellent to
miserable with statistical concentration somewhere be-
tween acceptable and good, and it became evident that the
objects in question must have the exclusive property of
being easily detectable independently of the quality of the
plates. In practice this implied that our "targets" should be
(a) very early-type stars, (b) very late-type stars, (c) stars
with particularly conspicuous spectra (for instance emis-
sion-line stars). At that time there was still a lot to be done
as far as the mapping and listing of these types of stars
vinced that the phenomenon itself was not necessarily
a unique astrophysical one, but merely had to do with
human perception. In other words: one reacts when two
almost equal spectra appear close together on the plate.
This concept could also be supported by means of
a comparison between the observed frequency of these
coincidence phenomena and the one calculated underthe
assumption that they were produced just by chance. The
result of this comparison showed a (not unexpected) clear
correlation between the relation between observed and
calculated number, as weil as the magnitude and angular
separation of the components. For stars brighter than
about m = 11 and with an angular separation less than ten
minutes of arc, there was an overwhelming excess of
observed coincidences, but the gradients were steep and
beyond m = 13 and separation = 12' the excess had
turned into a corresponding deficit. Thus: when the stars
(i. e. their spectra) are bright enough and appear sufficient-
Iy close to each other, one discovers the coincidence
immediately; otherwise one tends to ignore them. The
ultimate separation limit, of course, is set by the diameter
of the field of view in the inspection microscope.
Now the question was: how should we explain the
excess of coincidences that was actually observed? The
phenomenon itself was still an observational one and it
would be dangerous to state that there was only one
unique physical explanation behind it.
I-::'.~, 'Io'-._~-t _ .......... ,_·.r.-
. 't ~
ij....... ,....... ,.w' ~ ... ,f <I ,- .-
..,.~ .'-
.
~

. ..
... -.,.,
~

.were concerned and, in fact, we made adecent contribu-

tion to the first-approximation exploration of the Southern
Milky Way, particularly its very southernmost parts. My
most diligent co-workers at that time were Anita Sundman
and Birgitta Nordström.

Close Pairs of Similar Stars

During the inspection of the objective-prism plates, we

were immediately surprised by the extremely frequent
occurrence of two (sometimes more) spectra of (practical-
Iy) identical type and magnitude, situated so close together
that they formed a very conspicuous configuration for the
Fig. 1: Some examples of close pairs of objective-prism spectra of
nearly the same spectral class and magnitude. The three upper
pairs are all early type; the strong lines are the hydrogen (Balmer)
lines. The lower pair is of late (K) type and shows the prominent
G band at the centre.

8
 Our working hypothesis was that within a cluster (or
better: clustering) of stars there isan enhanced probability
·· ,
.;.
\ "', ... .. "
' ..
• ' • , .... \oe
. , ...... ... •
,.'
\ ' ...
.0' •

..
for accidental appearance of two equal stars close to the . .. '.'
"
" .... • ' •
. '.
I

"
,

same line of sight and that therefore a certain fraction of . ". ; ,"

. ~. ~ .~ .
.. .. ,
.." . .
'
the coincidences would probably belong to open clusters ,'0" •
.', .
~ . ,',.
or associations, many of which are too loose or poor to be . • • ........ • .,,', • • •
.
- .. t~ , ..
....
"
discovered directly by "conventional" methods. One might ".~. ...
.
'. '.'
..
'. ,.
'"
call these objects "cluster traitors". The coincidence
phenomenon is thus accidental but it facilitates the
,I'
·• •.. •~
,,",

•
';.'
• 0'>
.
~ '. '.'
,
. ".
detection of the cluster. One may thi nk of a pair of identical "
~. • . " ., .' <. . .'
I: .•- ~;, . . ".
twi ns in a gang wanted for cri me. If they appear far apart in " ;', .. ~ ':. . .,.:" ", :, -!' .•~.~.:
.... ~'. ' ...
the mob, their risk of being captured is appreciably smaller '.' '.
• ." ... ' o· ",' : ~ ,
0. 0, "'~ ~

than if they show themselves close together. ,;,.:". .~. '.~: •... . .~.' ,. "

~ . .' . ~ .~ ':.' >•. :. ':: ": ~ ..,:

. .•. .. ....
:"

...,: .. ... :,'.. . .. ~;:..'..' . ...•

'''.. • J I '. ••
~ f " .
.,
Observations at ESO , .";. .' ~.
.. ..... "; ~ :~. "
.'
Our first step was to check the relationship between the S'" ." ',: .. ! " , . * 0'
components of the candidate objects. For that purpose we " •. ,' . . . .' ,', ," . :. ' .: . •0' ....... :

used UBV photometry and slit spectra. The observations

were carried out at La Silla during 1969, 1973, 1974, and
1975 and the results showed that a considerable majority
..
: .. ~."
',":.
:: :.: '.
:
.''If
.
:. : . '
•.•
'. ..
'" t· . . • ;,
..••' ••...•. ,.' ..• ', .. " ;'.•
. . . ..
"
• ,',., '. .' 41
..:, "~':'"
. .' . - ~
'.
,. ...
..
'
..

.•
\,
..
.
of the coincidences were definitely situated so close to .: '. • • 'I,' .....

each other in space that there was very little doubt about
their mutual relationship. Next step was to study a selec-
tion of stars in their immediate neighbourhood in order to
discover other presumptive cluster members. This was
appreciably more difficult. The only "standard methods"
available were analysis of radial velocity or proper motion
data-under ideal conditions both. Unfortunately, none of
them could be used with any chance of success in this
case. Therefore we started desperate attempts to find
a physical criterion for cluster membership, preferably
a photometrically measurable one. Perhaps a certain
metallic peculiarity would be characteristic for a particular
cluster or association? We investigated the use of the
metal index m, in the Strämgren four-colour system for
discrimination of mutually associated members from stars
in the general field, and uvby observations were made at La
Silla between 1976 and 1979, together with UBV photo-
metry, of a large number of stars in the immediate sur-
rounding of some representative candidate objects.
Unfortunately it turned out that the metal index in
question was considerably more sensitive to temperature
and interstellar reddening than to subtle spectral details,
particularly for spectral types A and earlier which predomi-
nated in our material. For this reason we are now trying to
find a new criterion as a most urgent part of our present
research project.
Although we did not find any convenient, elegant, and
reliable discrimination method, we could, by means of
clumsy classical procedures, get a satisfactory answer to
our principal question concerning the suspected objects:
more than 80 % of the investigated objects are physically
real. A certain number turned out to be situated in or in
close connection with more or less ordinary clusters (some
of them already known or even weil known). A majority of
the other ones are very loose open clusters or clusterings,
weil in accordance with the working hypothesis. There is
also a significant number of very small and very poor
clusters which we prefer to call "microclusters" or "mini-
clusters", possibly some sort of cluster remnants. In a few
cases our candidate objects have more the appearance of
a binary or multiple system with extreme separation
between the components. It may be reasonable to assume
that such a configuration represents the ultimate state of
a cluster in dynamic decomposition. The last stars leaving
the cluster should be the most massive ones. The most
Fig. 2: A small part of the Milky Way in the constellation Caril)a.
Many groupings of stars are seen. The circle encloses a possible.
very loose cluster. A very close pair of stars (indicated with an
arrow) first drew attention to this area because of equal spectral
type and magnitude.
massive star and the second most massive one are
expected to have nearly the same magnitude and spectral
type.
How Many Clusters are there?
Next we come to the astrophysical implication of the
observed phenomenon. If we consider the spectral type
coincidence to be more or less accidental, we are forced to
believe that there are many loose stellar clusterings which
are less conspicuous and consequently remain undetect-
ed. It is even possible that we detect only a very small
minority of the clusterings and that the true number is
several orders of magnitude larger than the number of
ordinary clusters and associations catalogued thus far.
There still would be a majority of free stars without
dynamical connection to any particular cluster, but no
more an overwhelming one. The concept "general stellar
field" or "general stellar background" should in that case
be used with a certain caution. What we see when we
carelessly talk about this background field might as weil be
a puzzle created by a successive superposition of a mani-
fold of various clusterings and associations. Some ofthem
are visible; others disappear in the crowd. The denser and
richer a cluster is, the higher is its chance of being
detected.
If, on the other hand, stars of various magnitudes are
randomly distributed inspace there is also a high degree of
probability that a few of them accidentally will form
apparent groupings that are erroneously considered as
clusters. In other words: You don't see all real clusters in
your Milky Way region and all clusters you see are not real.
And there is no reason to believe that the two processes
cancel out.
With respect to the general structure of the Milky Way,
we can make two different models, each one representing

9
an extreme case. To begin with we forget the time parame-
ter. In the first one the stars are completely homogeneously
distributed-except for some random groupings. In the
second one there are no "free" stars at all, only clusters
and associations of clusters in some sort of hierarchic
organization. Observations tell us that none of the models
is correct but that the real situation is somewhere in
between these extreme cases. The natural question then
is: Where in between? Here we must introduce the time
parameter; the relation between "free" and "bound" stars
is definitely correlated to the stage of evolution of the
Galaxy-unless we believe that the processes of creation
and disintegration of c1usterings are in perfect equilibrium.
Personally I do not think that it would be realistic to use
the number of cluster(ing)s as a criterion of the age of
a galaxy but, more probably, it could be used in an
empirical check of the theories for the dynamical stability
of these clusters and that is interesting enough.
Fig. 3: The distribution of known open clusters in 60 0 intervals of
galactic longitude. The small rectangle indicates the area investi-
gated by the author and his collaborators at the Stockholm
Observatory.
Discovery of New Wolf-Rayet Stars in the Magellanic Clouds
J. Breysacher and M. Azzopardi
As a result of a thorough seareh with the ESO GPO
astrograph, the number of known Wolf-Rayet
stars in the Small Magellanie Cloud has just been
doubled (from 4 to 8). Drs. Jaeques Breysaeher
(ESO) and Mare Azzopardi (Observatoire de Tou-
louse, Franee) also diseovered 17 new WR stars in
the Large Magellanie Cloud. Slit speetra of these
stars have been obtained with the 3.6 m teleseope
and there is an indieation of a signifieant differ-
enee between the WR stars in the Clouds and
those in our own galaxy.
The Magellanic Clouds offer the possi bi lity to study objects
of various classes which are at the same distance from uso
It is known that some notable differences exist between the
stellar populations of the two Clouds, and the WOlf-Rayet
(WR) stars do not seem to be an exception to the rule.
However, before any comparative study can be under-
taken, it is fi rst necessary to make sure that the detection of
WR stars in both Clouds is as complete as possible.
The Objective-Prism Search
A systematic search for this kind of star was carried out in
October 1977, in March, and in November 1978 with the
10
I I. a_ 60°
n (.0 -120
llJ 1 ~() -let,)
IV IBe -240
V 240 -300 -
VI 300 -360
200 -
...---
N
r-~
f---
~
100
~ II III IV v VI
ESO 40 cm Objective-Prism Astrograph using an inferfe-
rence filter centred at A 4650 which has a passband of
120 A wide. WR stars show up strongly in this spectral
region due to the emission, mainly from either A 4650 C 111
(WC) or A 4686 He 11 (WN). This detection technique
enabled us to study very crowded regions by reducing the
background fog and the length of the spectra, i. e. the
number of overlapping images.
Figure 1 reproduces an LMC survey plate. The field has
85' in diameter. The limiting magnitudeofthesurveyism pg
= 16.5 for the Small Cloud and m pg = 17.5 for the greater
part of the Large Cloud. But due to the poor sensitivity of
some of the lIa-O plates used, for a few LMC fields only the
continuum of 16.0 m pg stars was reached. The B mag ni-
tudes of the WR stars were determined from astrographic
plates taken after removi ng the prisms of the Objective-
Prism Astrograph, in combination with a Schott GG385
filter. In order to get an accurate classification of the newly
discovered WR stars, slit spectra were obtained for all of
them with the Boiler and Chivens Cassegrain spectrograph
equipped with eitherthe Carnegie image-tube orthe Image
Dissector Scanner at the ESO 3.6 m telescope.
SMC
In the Small Magellanic Cloud, 4 new WR stars of the WN
type (12.9:::; m pg :::; 15.3) were identified (Azzopardi and
Breysacher, 1979a) increasing to 8 the number of known
WR stars in this system. Considering the distribution
among the different WR subclasses, it is remarkable that in
the SMC only subclasses W3 to WN 4,5 are present with, in LMC
the WC sequence, one doubtfully (Breysacher and Wester-
lund, 1978) extreme Wolf-Rayet of type WC4. This has In the Large Magellanic Cloud the present survey led to the
possibly something to do with the general metal deficiency detection of 17 new WR stars of the WN type (11.9 :::::; m pg
of the SMC and one of us (J. B.) is now studying this point. :::::; 16.4). 13 of which are in the region of and to the west of
Adopting the absorption-free distance modulus of 19.2 for the 30 Doradus nebula (Azzopardi and Breysacher, 1979b,
the SMC, we come to the result that the 3 faintest SMC WR 1979c). 101 WR stars are now known in the LMC; the
stars, also binaries, have absolute magnitudes which are corresponding figure for the Galaxy is 154.
hardly compatible with the existing absolute magnitude With an LMC distance modulus of 18.5 the absolute
calibrations for WR and OB stars. These 3 WR binaries magnitudes obtained for 8 WR stars (m pg > 15) are signifi-
located in the same south-west region of the SMC tend to cantly fainter than the values given by Smith (1973) for the
confirm that the extension in depth of the Small Cloud is corresponding subclasses. Since the Large Cloud is
rather large, as previously suggested by Hindman (1967) generally considered as a system which is seen almost
from 21-cm radio observations. "face on", local stronger absorption is possible but it may

,. I ,I
, ,
, , "
I,
',I 'I
, ,I ' I ,,
I"

"
"
I
1
'I "

I,'
I
" 1
I
,, ,, I
,
, , I , ,I

, . ,
I 1 '
,
,.1
. I
,,I,
, "
I' "
I "
.,
I
11I
I,
.,
1I, ' I

I'
, I
I
,,.1"
, 1 I ,, I
,
1 I
",
,• ",
, I' ," ,li

, , , \' , I

I,
I I, I I, 1
1
I
1 ,, h 'I
"
1
,I,
1_/
"
, I" "

,,
I ,

" I
, I' I, I
'I
,I , I.',
\'
,',"I
,'
I,
I , 1 , I, , ,'I~," "1 , "
I I

,/'1 1 I ,I , I" , ,
I'
I ,
1
'I' "1 ,
I
, I'
'I
"
",, , " I " if
"1
,, I '
I
I
,t ~,
, , 11 '
I II'~
I
,
'I I
11, I' 'li,
' .~
, I.
, I
"
I' I
,I
~I"
I
'I I ' '" I1 ,
, • ' 11 ' [, I
"1 , 1\ I~ '11 I" IJ, ' I I' , " I, \, 1'1 1,1
I'

"I~'"""
"
"
,
I I
11 ,~ "I fll'
I ',I '" I 11 I " ' ", ,, I'
\ 'I 'I ' ,, " ",'", 11 11 I \1,,' I
,
I '
, I I ' ,I , '\ 1 'I 'li
/ I
"
II 1\' I I ,I
, I'"
. , , I I "
,'I ,I ' I
I ,I' ,
',(
,I
,\ 1
"I ,', "' I' ", 1 ,,11
I, I" \ 'li I' "I ," I' t' I I
I, ,
,
11
, I , ,
.
"
I'
d,11 ~"I"'I, I 1 I , "

"
,,
, , ~ , '11" "
" '
" '1111
"
'I
,\,1,1,11'
I"
\
1'~
' I"
,' I1\""~ I t ,I' '
" I I
, I, I I , I'
.1 1
, I I I . \1 ~ 11 1 , I I I I I .. I , I'
,I 1
'
11

",
' '\ ", " " " " , I' "I, .\
\ ~'. ,', ~ "I 'I I', \" " I If I", ,1'111 "
~ I " 1 1, 11 I'" I I ' ,' ,
, ,.'
l'~ 1 ' I,
'I , I, \ I "
,,
}
'I' 11 .. i I ,,'\ I, /
"r"11 '~I I Ilf,jIr~\"I~IIIIII~11
I• "
1 "
-, I t
tI I ' " , I

,I.' , , I 'li
"
I " I I, I
I, I I ,
Ij ",'I
,

.
, ,
111,
I
, "I
,
, '
I I'
"li, "
l'l'
I ~ ,
,'
,
,
I'
11
I'

I
I

I
, "
I,
I
'I. "
,

' I
I'
,

"
'1,'1 1 1
'1,,1
, I11 - ,
"
',li
I

Fig. 1: A 6-hour exposure on 11 a-O nitrogen baked plate in the 30 Ooradus region obtained in Oetober 1977 by M. Azzopardi with the
Objeetive-Prism Astrograph and the A 4650 filter. Wolf-Rayet stars with their emission feature are easily reeognizable. Two of them are
indieated byarrows.

11
weil be that these WR stars really have lower intrinsic
luminosities than those stars of similar types previously
observed in the LMC. This is now being investigated.
The census of the WR population in the Magellanic
Clouds can now probably be considered as quite com-
plete, except, maybe for subclass WC 5 which has possibly
escaped our detection due to the technique employed: the
width of the A. 4650 emission feature is comparable to the
filter passband in this case.
References:
Azzopardi, M. and Breysacher, J.: 1979a, Astron. Astrophys. (in
press).
Azzopardi, M. and Breysacher, J.: 1979b, Astron. Astrophys. (in
press).
Azzopardi, M. and Breysacher, J.: 1979c, Submitted to Astron.
Astrophys. Suppl.
Breysacher, J. and Westerlund, B. E.: 1978, Astron. Astrophys. 67,
261.
Hindman, J. V., 1967, Australian J. Phys.20, 147.
Smith, L. F. 1973, lAU Symposium, N°49, 15.
Instrumentation Schedule
This is the up-dated time schedule for the major instruments
which are being developed at ESO in Geneva for use on the 3.6 m
telescope. See also Messenger No. 15, p. 10.
Triplet Adaptor (M. Tarenghi, M. Ziebell). Target date: Sept. 1979.
The components are:
- two 3-lens correctors for prime focus
- an adaptor with tv for acquisition and guiding
- a remote-controlled shutter and changer for 4 filters
- a remote-controlled changer for 8 plates (3 magazines); plate
size is 240 x 240 mm.
For more details see MessengerNo.16.
4 cm McMullan Camera (W. Richter). Target date: October 1979.
- Electronographic camera as developed by McMulian. Can be
used behind triplet adaptor in prime focus.
Couda Echelle Scanner (CES) (0. Enard, J. Andersen [Copenha-
gen], A. Danks). Target date: mid 1980.
- instrument to record very high resolution digital spectra (up to
100,000) on a 1876-channel-DIGICON detector. Double-pass
scanning mode permitting calibrations on bright objects with
very clean instrumental profile.
For more details see Messenger Nr. 11.
Couda Auxiliary Telescope (CAT) (T. Andersen, M. Dennefeld).
Target date: mid 1980.
- 1.5 m spectroscopic telescope feeding CES of the 3.6 m
telescope. Three-mirror alt-alt telescope with f/120 (f/32 after
focal reducer). Dall-Kirkham optics with spherical secondary.
Direct drive servos without gear.
For more details see Messenger No. 10.
Infrared Top-End (R. Grip, P. Salinari). Target date: mid 1980.
- Wobbling secondary mirror with f/35 in Cassegrain focus, new
telescope top-ring which puts radiating material away from
light beam.
For more details see Messenger Nr. 13.
Cassegrain Echelle Spectrograph (CASPEC) (M. le Luyer, J.
Melnick). Target date: end 1980.
- Instrument with resolution of 15,000, 30,000 and 60,000 with an
SEC-Vidicon detector. Data-reduction process not yet defined
in detail.
More details are published on page 27 in this Messenger.
Compared to the schedule which was published three months
ago, the target date for the Triplet Adaptor has changed from
before to after the holiday period.
12
NEWS and NOTES
The 100th Anniversary
of the Birth of Bernhard Schmidt
Fig. 1: Bernhard Schmidt (1879-1935).
The inventor of the so-ca lied "coma-free telescope" was born
a hundred years ago, on March 30, 1879, as the son of a poor
fisherman on the island of Nargen in the Baltic Sea near Reval in
Estonia. Already as a child he experimented scientifically, and he
lost his right arm, due to an explosion in his primitive laboratory.
In 1901 he registered as a student of engineering sciences at the
Technical High School at Mittweida in Germany. Very soon,'
however, he gave up his regular studies and became independent
as designer and constructor of small o.ptical elements for
amateurs. He, himself an outstanding amateur astronomer, is
known as one of the first explorers of Nova Persei 1901.
Due to the high quality of his products and the deeply founded
knowledge in practical optics, he soon (1904-1913) became an
independent collaborator to the Astrophysical Observatory at
Potsdam under K. Schwarzschild and later at the Hamburg
Observatory at Bergedorf under R. Schorr.
During a long travel to the solar eclipse at Manila he accom-
panied W. Baade. Maybe inspired by him, he conceived the
famous telescope, which in 1930 got its final shape in the "Original
Hamburg Schmidt Telescope". This first Schmidt with a free
aperture of 36 cm was areal optical sensation. With a hitherto
unbelievable F ratio of 1: 1.75 it covered a field of 15 degrees of
diameter, completely free of all optical aberrations, except field
curvature. Shortly after Bernhard Schmidt's sudden death in 1935
the Schmidt telescope started its triumphal procession through-
out the astronomical world. There is a straight line from the
Original Schmidt to the big Hamburg Schmidt and finally to the
ESO-Schmidt telescope on La Silla.
On the occasion of his centenary the Hamburg Observatory, in
cooperation with the Astronomische Gesellschaft, organized an
international meeting of observers with modern Schmidt tele-
scopes, showing the ever-growing importance of the Schmidt
telescope as an instrument especially suitable for all kinds of sky
surveys.
On March 30, 1979 a small Bernhard Schmidt Museum on the
site of the Hamburg Observatory was inaugurated where a number
of optical elements and tools made by his own hands have been
collected. Most important of all, the original handwritten manu-
scripts, hitherto unknown, could be shown for the first time to the
public, due to a generous gift of the heirs of R. Schorr. They show
that Bernhard Schmidt very carefully studied the theoretical
conditions of his problem before he set out to realize the first
Schmidt telescope. A. Sehr
Fig.2: The Original Schmidt Telescope at the Hamburg Observa-
tory; f/1.75, 15° field.
Astronomy in Europe
Professional astronomers sometimes receive letters that begin
like this: "Dear Mr.... , I am 16 years old and very interested in
astronomy. I should like to know how I can become an astro-
nomer.
All astronomers are happy to see examples of deep devotion to
the science they like themselves. So most of them answer such
letters with kind words of encouragement and an explanation
about the study of astronomy at the national universities. But there
is one point-and probably the most important of all-the
prospects for a successful career in astronomy, that is very
difficult to answer. Few sciences have avoided the effects of the
present tightening of financial resources, and the number of
positions that are available at universities and other research
institutions is small. So the letter to the young astronomer-in-spe
will have to tell him that he has little hope of ever getting a job in
astronomy-unless, of course, he is very brilliant.
But how bad is really the situation? A partial answer is given by
areport "A study of man power in astronomy in the countries
represented in the European Science Foundation", that was
published by ESF last year. This report, for the first time, assigns
quantitative figures to the shortage of posts. It also includes
a number of interesting findings concerning the European astro-
nomical community.
The total number of astronomers in the countries that were
studied (16 in all, including the 6 ESO countries) is about 2,400. In
this connection, an "astronomer" is defined as somebody with
a PhD or equivalent competence, working actively as a scientist in
a field of astronomy. The corresponding figure in the USA is about
1,400. The European age distribution is peaked near 37 years and
about 50 % are in the age group 30-40 years. The number of posts
in astronomy (in Europe) that will become vacant due to normal
retirement during the forthcoming 15 years is about 350. When
taking into account the number of new posts that are likely to be
created, an optimistic figure is about 50 new posts per year. This
should be compared with the 220 new astronomy PhD's that are
produced every year. It is therefore clear that only about 20 % of
those who obtain a degree in astronomy will also obtain a perma-
nent position within astronomy.
There is one positive aspect, though: very few people with an
astronomy degree are actually without work. It appears that
astronomy as such includes so many valuable assets that
a candidate will have little difficulty in finding jobs in related fields.
Many astronomers teach physics and mathematics in schools and
others are employed in industry. The knowledge of practical work
in astronomical instrumentation and in particular experience with
computers is of value.
Funding of astronomical research has increased by about 40 %
(in real terms) between 1970 and 1976, proving the importance of
astronomy as a fundamental research discipline in the eyes of
budgetary commissions in the various countries. And astronomy
is obviously able to attract the best people everywhere. Therefore
the present situation is clearly very promising for the science as
such, but less so for its many admirers. Fortunately, astronomy
can be enjoyed equally weil by the scientist in the prime focus cage
of the largest telescope in the world and the amateur in his
backyard with his home-made reflector. This is a great advantage
in comparison with most other sciences, also if we consider that
the amateurs still play an important role, first by finding novae and
comets and in general by monitoring the skies from all over the
world. Perhaps the Olympic motto is also valid in astronomy: "The
most important is not to win (i.e. to become a full-time profes-
sional), but to participate!"
Echelle Spectrograms
An informal ESO workshop to discuss data-reduction techniques
forechelle spectrograms was held in Geneva on March 1-2,1979.
Astronomers and engineers from all ESO countries already
having considerable experience in the treatment of echellograms
gathered at ESO-TP to share their experience and to discuss the
various difficulties inherent in the reduction of echelle data.
The techniques being used at present to reduce data from the La
Silla and Haute-Provence spectrographs, and from IUE and the
Utrecht Ultraviolet balloon experiment were discussed in great
detail with the aim of familiarizing the ESO astronomers, engi-
neers and computer programmers responsible for the design and
construction of the ESO 3.6 m Cassegrain echelle spectrograph
(cf. p. 27) with the results of other European groups which operate
similar instruments.
The main conclusion from this workshop may be summarized by
saying that the extraction of astrophysically useful information
from echellograms may be difficult but is by no means impossible!
The workshop was a good starting point for the ESO team
responsible for the operation and data reduction of CASPEC. The
software is being written in parallel with the hardware construc-
tion and should be completed and tested weil ahead of the time
when CASPEC is put in operation. J. Melnick
Proceedings of the ESA/ESO Workshop
on Astronomical Uses of the Space
Telescope
The Proceedings of this workshop have now been edited and
will be available in print by end of May 1979.
The price for the 450-page volume is SFr. 40.- (in Europe)
and US$ 20.- (elsewhere), including postage. Please send your
order to:
European Southern Observatory
c/o CERN
Attn. Miss M. Carvalho
CH-1211 Geneva 23
13
0.5 Arcsecond Images with the Danish 1.5 m Telescope
on La Silla!
J. Andersen and B. Niss
The 2.4 m Space Telescope will achieve (J'1
resolution in 1984. But what is the best possible
angular resolution from a ground-based observa-
tory? Recently, fantastic long-exposure plates
were obtained with the Danish 1.5 m telescope at
La Silla, proving at the same time the excellent
performance of this new telescope and the quality
of the ESO site. Drs. Johannes Andersen and
Birger Niss from the Copenhagen Observatory,
Denmark, tell the exciting story.
In the last issue of the Messenger, the general features as
weil as the optical alignment, commissioning, and initial
performance of the Oanish 1.5 m telescope on La Silla
were described. The conclusion, based mainly on labora-
tory tests and the accuracy achieved in the alignment, was
that the image quality until then had been entirely limited
by seeing, but confidence was expressed that the tele-
scope would be "able to take advantage of even the nights
of very best seeing".
Such prophecies are not uncommon in articles describ-
ing new telescopes. They are usually met with a benevolent
scepticism of seasoned observers, who know by experi-
ence all the good excuses why the theoretically predicted
image quality is (almost) never experienced in practice:
L-.J
,"
Fig. 1: Enlargement from a 1-hour exposure on lIIa-J emulsion
behind a GG385 filter, showing stars in the globular cluster
NGC 3201. The images are perfectiy round and measure about
0.5 arcsecond in diameter. Danish 1.5 m telescope; observer Dr.
B. Niss, March 7, 1979. Average zenith distance 18°.
14
Seeing, whether external or internal in the dome or
telescope tube, and imperfect optical alignment, mechani-
cal stability, and/or guiding, all combine to make images of
one arcsecond or slightly less the best one hopes for in
longer exposures, even if theoretical resolution is half that
figure or better.
The sanguine predictions for the Oanish 1.5 m telescope
were, however, confirmed before the ink on them was
dry-with one startling reservation as will be discussed at
the end of this note. In early March this year, one of us
(8. N.) was continuing the observing programme in globu-
lar clusters described in MessengerNo.10, p.14. Although
the telescope was (and still is) in the testing phase, good
cooperation from the equipment combined with aspeIl of
excellent atmospheric conditions to produce a superb
collection of plates. The image sizes range from 1"through
several plates of 0:7-0:'6 to the best one, a one-hour
exposure on 111 a-J emulsion of NGC 3201, which shows
images nicely circular-and of diameter 0:'5 (30 microns)
as measured on a projection micrometer! The figures show
a reproduction of this plate and a POS scan through one of
the images.
This was an almost unbelievable result (J. A. was in fact
only convinced by his own eyes looking through the
micrometer eyepiece!). As mentioned in the previous
article, the mirror acceptance tests indicated a geometrical
energy concentration of 80 % in 0~'45, to wh ich must be
added the diffraction disk ofO:'2-and you already have the
observed diameters! In fact, had these images been taken
in a laboratory vacuum test tank, they would have been
considered a most gratifying confirmation of the more
indirect test methods. Obtaining such images in a long
exposure with areal, moving telescope in a real dome and
equally real atmosphere is an entirely different matter;
however, not the least if one keeps in mi nd that asymme-
tries of 0:'1-0:'2 would have been plainly visible! This leads
us to several pleasant conclusions:
- The optical test results supplied by Grubb Parsons
were probably even on the conservative side;
The optical alignment was in fact done to better than
0:'1 of coma, as previously described, and it remained
intact after four months of operation;
The telescope tube and drives are of excellent mecha-
nical quality;
The autoguider and control system achieved a guiding
accuracy of about 0:'1 as specified, and, last but not
least,
Seeing, external plus internal, was significantly better
than 0:'5.
We leave the many possibilities offered by such images to
the reader's imagination, but a quick comparison with
previous electronographic work indicates that had one of
our McMulian cameras been on the telescope that night,
we would have been able to detect and measure stars of
magnitude between 26 and 27! We do not suggest that
such nights are the rule, even on La Silla, but nor do they
belong entirely in the realm of dreams.
If one insists in being ungrateful, than it should be said
that our mirrors, which we always considered excellent,
did not in the end live up to those "nights of very best
seeing". Rather unexpectedly, the resolution seems ulti-
mately to be limited by residual optical (mostly zonal)
aberrations, even at this very low level. It will be interesting
to see whether the CAT optics, made by the same
manufacturer under even tighter specifications, will pro-
duce still better images under optimum conditions.
Being far from ungrateful, however, we wish to conclude
by paying tribute once again to those responsible for this
achievement: to the firm of Grubb Parsons for their
outstanding optical and mechanical craftsmanship, to the
ESO Optics Group for their invaluable help in testing and
aligning the optics, and to the ESO Controls Group and the
workshops of Copenhagen University Observatory for the
successfu I combination of control system and autoguider.
The Problem of Star Formation-and what
Ten Nights of Sub-millimetre Observations
with the VLT Could Contribute to its Solution
P. G. Mezger
Astronomical observations are regularly carried
out over the whole electromagnetic spectrum,
from y-rays to low-frequency radio waves. There
are few unexplored "holes", but one of these-in
the neighbourhood of 1 mm-is exactly where we
expect most of the radiation from stars during
their early stages of formation. The VL T would be
ideally suited for ground-based observations in
the sub-millimetre range, because of its large
surface and good angular resolution. Dr. Peter
Mezger of the Max Planck Institute for Radio-
astronomy in Bann explains how the VL T can
make a very important contribution to the study of
stellar formation.
Sub-millimetre Observations, Star Formation and
the VLT
The transformation of gas into condensed objects, either
ordinary main-sequence stars with masses -0.1-100 MO
or perhaps also much heavier supermassive stars, is one of
the most fundamental processes in the Universe. Star
formation plays a leading role in the formation of galaxies,
in the chemical evolution of the interstellar matter (i.e. its
enrichment with elements heavier than 'He) and may weil
be related to some of the phenomena associated with radio
galaxies and quasars.
In spite of a wealth of radio and IR observations related to
both dense molecular clouds (out of which protostars
form) and pre-main-sequence evolutionary stages of mas-
sive stars, the basic process of the formation of protostars
out of the interstellar matter is far from being understood,
even in a qualitative way. The reason is that the formation
of protostars occurs at very low temperatures of the
1 ..
Fig. 2: A scan /hrough one of /he images in figure 1, made wi/h /he
POS microdensi/ome/er a/ /he Lund Observa/ory /hrough
a 10 x 10 micron square aper/ure.
interstellar gas (typically -10 K) and that the outer shell of
the contracting protostar remains at such low tempera-
tures until nuclear burning starts at its centre. Thus the
Planck curve for 10 K (shown as dash-dotted curve in
Figure 1) is an upper limit for the intensity of both
continuum and line TE radiation emitted by dense molecu-
lar clouds and protostars in their early evolutionary stages.
This curve peaks at -500 /-lm (= 0.5 mm).ln Figure 1 isaiso
shown the transparency (heavy curve) of the atmosphere
for an amount of 1.3 mm of precipitable water, conditions
as they prevail at an altitude of -3,000 m for about 30 % of
the clear nights. One recognizes a number of atmospheric
windows whose transparency decreases with decreasing
wavelength. Below -300 /-lm the atmospQere is practically
opaque. The wavelength range between 1.8 mm and
300/-lm, although accessible for ground-based observa-
tions with a telescope placed at a very high and dry site, is
largely unexplored. This is due to both a lack of sensitive
radiometers and of radio telescopes with a sufficient
surface accuracy of its reflector.
Promising developments of both coherent radiometers
(for spectroscopy) and incoherent radiometers (bolome-
ters for broadband continuum observations) for the sub-
millimetre wavelength range are in progress in various
laboratories in Europe and the USo But even the second
generation of mm-telescopes, now being planned or under
construction, are only marginally usableforsub-millimetre
observations. The reason is that the quality of a telescope
for coherent detection is determi ned by the rms deviation
of its reflector surface from a best-fit paraboloid, and this
in turn is determined by the surface accuracy of the
reflector panels, the accuracy with wh ich these panels can
be adjusted, and by the design of the reflector back-up
structure. Most mm-telescopes in operation today have
rms deviations 0 ~ 100 /-lm. For the new large mm-telesco-
pes one anticipates rms deviations in the range 90 ~ a//-lm
~ 50, which degrade the telescope characteristics (such as
gain, aperture, and beam efficiency) according to exp
15
 FREOUENCY (GHz)
Z
10 Z 5 10 Z 5 1d Z 5 104
...."1'"=F=-=-=-:::-_r-_....-_'I-==:.-":r:.-=----r--_-):r.:,..,.......!,.,-r'I------,---rI----,--,-'I---r--.--r-r'""l
1r--'--"'Ioo:::s;;;:::::;;;;oo;;:;;;;:~: -r-.....--,rr=-=.~·-::::-:_r-._,..A
....---........ / .... 1........... ~ / ·~.---Bv/ BVM
"
", '
,, ~ ,;<.I '\,
.\ , ' 10 K BLACK BOOY
"\ "" . ,
\ 1/ \ \ \
\
\ \ ''\ \ \
"\
PRECI P
HZ O(MM)
\ .\
\
\ !
\ ! \,\ \\ .\ \\ -1.3
/
, \ \
\/· \ \
\ \ \. \
\
-

/
X, \ ',.
\ \,
. 100\ 50 \ Z5 \ ,10 ... (J ( 11 M)

/
/
/,
\
\
,
\r.

\
'\
\.
\\ \.
\
\
\\
/ \ \ \
\ .
",""
\
,/ [\ \ \ \ \
// \ \ \\ '\.
'.
,/

,...... ...... ~\',',

o L.----II l--~.-_.--'---'-"---'----L--1--lI---L.....L...-_..L..I.-
.=J..: I...L....LJL'..oL:....J-------I..1I.....-...l..&-..::..::...----LI_'....,..l:::~.>_.:·
::::..,._J.L...J........L......J.._.... '__.lI_l_...J
...... .L.:::
....

ZO 10 5 Z 1 .5 .Z .1 .05
WAVELENGTH (M M)
Fig. 1: Transparency of the atmosphere (heavy curve) for an amount of 1.3 mm of precipitable water. Relative change of gain and
efficiency of a telescope used for coherent detection (dashed curves). Curve parameter is the rms reflector deviation 0 from a best-fit
paraboloid. And (dash-doffed curve) normalized Planck radiation curve for 10 K. (Adapted from Leighton, 1978, Final Technical Report
NSF Grant AST 73-04908).

{-(4Jt on.)2}. This function is shown in Figure 1 for know that giant molecular clouds have masses 105-10. Mo,
different values of 0 as dashed curves. Telescopes to be mean densities of 103-10' cm 3 and kinetic gas tempera-
used in the sub-millimetre range A ~ 300 ~m should have tures ~ 10 K. And it appears that at densities ~ 10' cm- 3 gas
rms deviations 0 ::s 10 - 20 ~m. Today, such accu racies are and dust are in thermodynamical equilibrium. But what
only attained by optical reflectors and this explains why at initiates (or inhibits) star formation in these clouds we do
present observations in the sub-millimetre range are not know. We have learned from recent model calculations
exclusively done with large optical telescopes, mostly at that (at least the more massive) stars form by accretion. At
twilight and during moonlight nights. Considering these the centre of a contracting protostar, density and tempera-
facts und the observing interests of me and some of my ture become high enough for hydrogen burning. An
associates at the MPlfR it is obvious that we would use VLT embryo star forms, which grows by infall from the outer
observing time for an investigation of the early phases of layers of the protostar und therefore evolves up the main
protostars. Within ten "nights" at the VLT, sub-millimetre sequence. After the embryo star stops accreting, a shell of
observers with improved coherent and incoherent radio- gas and dust is left behind, which is visible first as a FIR
meters should be able to gain insight into some ofthe basic source and subsequently as a compact H 1I region. The
processes of star formation. I have deliberately put sequence of observable stages of the protostellar shell,
"nights" in quotation marks, since sub-millimetre ob- after hydrogen burning has started at the centre of the
servers would probably always request only such observ- embryo star, is shown in Figure 2. But practically nothing is
ing time on a VLT which could not-or only marginally- known about the earliest evolutionary stages of protostars,
be used by optical observers. when these objects shou Id appear as dense but isothermal
condensations in cool molecular clouds.
The situation becomes even worse if we turn to the
Wh at 00 We Know about Star Formation? formation of lower-mass stars (i. e. stars with masses less
than a few solar masses) which account for the bulk of the
From recent radio and infrared observations we have mass of stars. While both high-mass stars and lower-mass
learned quite a lot about the pre-main-sequence evolution stars form out of massive and dense clouds of interstellar
of massive OB stars. We know that these stars form out of matter, the former appear to form predominantly in the
dense clouds of interstellar gas where practically all main spiral arms, the latter appear to form predominantly
hydrogen is in molecular form. H2 has no observable radio in the interarm region (observable in some cases as T-Tauri
transitions, but since it is the most abundant collision associations). What determines the stellar birthrate func-
partner its density can be crudely estimated from the tion, why is it easier for nature to form low-mass stars than
intensities of collisionally excited transitions of moleeules massive stars (in contradiction to what the Jeans criterium
such as CO, CS, CN or HCN. From such observations we tries to tell us), and which fraction of mass ends up in

16
substellar objects with masses =s 0.07 MO? Answers to
these questions will probably only come from sub-milli-
metre and FIR observations.
Emission from Dust, from Moleeules and from
Atoms at Sub-millimetre Wavelengths
Let us first consider the quasi-thermal emission from cool
dust. At sub-millimetre wavelengths ~ 300 I-lm even forthe
giant molecular clouds the average dust optical depth is
smalI, so that protostars which form deep in these clouds,
due to their higher densities and hence optical depths,
should be observable as emission centres or "hot spots".
Hildeband and his colleagues of the University of Chicago,
using optical telescopes at Cerro Tololo and Mauna Kea,
have actually observed thermal dust emission from cool
molecular clouds and from the globule Barnard 335. For
the latter the colour temperature of the dust was found to
be - 8 K. This type of observations with telescopes of high
angular resolution (a 25 m VLT at A. 500 I-lm will have
a HPBW of -5 arc sec) should be a very powerful tool for
the investigation of the earliest protostellar stages. At
present, one uses composite Ge-bolometers which are
FREE FALL CONTRACTION
cooled to temperatures of - 1.5 K, attainable with pumped
liquid 4He. However, temperatures as low as - 0.3 K can be
attained by using 3He as a coolant, and this should increase
the sensitivity of future bolometers by at least an order of
magnitude.
The optical depth of dust is independent of temperature
and varies only slowly with frequency. The spectral shape
of optically thin dust radiation thus is still very similar to
that of a Planck curve. On the other hand the optical depth
of the rotational transition of a molecule increases with
Tex 2, but decreases rapidly with increasing rotational
quantum number J or decreasing wavelength once E(J)
> kT. Figure 3 shows the intensity distribution of rotational
lines of the CO molecule, computed for Tex = 10 K; curve
parameter is L eo (J = 1~O), the optical depth of the lowest
transition. Even for L eo (1~0) = 10-100 (which may be
typical for dense molecular clouds) the lines with J ~ 4-5
become optically thin and thus allow observations of
condensations in the cloud. Again, the high angular
resolution of the VLT should allow to observe dense
condensations inside the optically thin clouds, wh ich
I would expect to be the first evolutionary stages of
protostars. Equally important is the fact that the analysis of
several (optically thin) rotational transitions of one mole-
cule leads to a much more accurate determination of the
physical state of a molecular cloud than does the currently
applied method of observing one handy transition at mm
wavelengths A.~2 mm of different molecules with different
dipole moments. Plambeck and Williams, for example,

(\1O~ 100~ lmm IR SOURCE

compared density determinations from the J = 1~0 and
2~ 1 transition of the isotopic "CO, with corresponding
densities derived from one transition of CS, CN and CHN,
respectively. They found that the latter method overesti-
mates densities by typically a factor of 10! This is one
-S·10 4 yr reason why the sub-millimetre range with its possibility to
observe several rotational transitions of one molecule, is
so important for the astrophysical application of molecular
spectroscopy. Other reasons are that several molecules,
such as the possibly very important species of hydrides,
>-
have such low moments of inertia that their rotational
I-
V>
ZAMS spectra only start at wavelengths < 1 mm. Observations of
=
LU
=
><
COMPACT (ibl 2.6 1.3 087 065 0.52 01,3 037 033 029 A/mm
=
--' Hll REGION -l10 4 yr
l..L..
OH MASER? 100 1 I1
v v mox
/'1
,, \
\
,/ \
\
10~ 100~ lmm lem 0]5
\
\
\
\
COMPACT (dbl \
\
\
Hll REGION - S'10 5 yr \10
\
=
\

~-l~m-m---'le-m-o~:7~'-ftAl Y t
\
\
0.50 / \L l (J;j+OI \,
,/ \;1 \
\\ \\
1------...:--1...L
O-O V151 BlE \ ~
\
\
\
\
Rolotlonsspektrum von
6 \ \
12CO
Hll REGION -S'10 yr 025 ,
\
\
\ I., 10 K
1
\ \

MS TURN OFF
\ \
\
\
\
\

Fig. 2: Observable spectra of an evolving massive protostar. Time

scales on the right side correspond to a main-sequence star of
17 M0 . During the free-fall contraction the protostar should be
observable as a 10 K black body (see Fig. 1). A compact radio H 11
region forms after the accretion has stopped and the star has
attained the ZAMS. The maser, IR and Radio stage of an 0 star
together last for-5-10 % of its MS life time (Mezger, 1978, In fra red
Astronomy [G. Setti and G. G. Fazio, eds.j D. Reidel Publ. Co. p. 1.).
Fig. 3: Intensities of rotationallines H (J-1) of the "CO molecule
computed by Schmid-Burgk tor a plane-parallel gas layer in TE at
10 K. Curve parameter L eo (J = 1 0) is the optical depth for the
lowest rotational transition. The curve 'co 00 corresponds to the
Planck curve. Note that the width of rotational lines increases
approximately proportional with J.

17
hydrides may in fact playa key role for our understanding
of interstellar chemistry. Of special interest for astrophy-
sics are transitions of atoms and atomic ions since they
allow abundance determinations. Furthermore, fine struc-
ture lines of atoms and ions as Co and C' provide part of the
cooling for interstellar clouds. The A 610/-lm line of
Co should be of special interest for the physics of dense
interstellar clouds and may be observable with a ground-
based telescope. However, only heterodyne radiometers
will provide sufficient sensitivity for sub-millimetre spec-
troscopy. As mentioned before, intensive development
work on this type of radiometer is in progress in several
laboratories in Europe and USA:
European Astronomers Discuss the Use
of the Space Telescope (continued)
With the publication of the Proceedings of the ESA/ESO
Workshop on "Astronomical Uses of the Space Tele-
scope", held in Geneva on February 12-14, 1979, it is now
possible to better judge the interest of the European
astronomical community in the Space Telescope. The
Proceedings of course only represent the "official" part of
the meeting. There were also lively discussions among the
186 participants and it is impossible to write them all down!
SECONIlARY
MIRROR
~
SECONIlARY RAFFLE
CENTRAL RAFFLE
FOCAL PLANE
Fig. 1: Tl7e Space Telescope.
18
There are other, more speculative, objects to be ob-
served in the sub-millimetre range, such as the redshifted
dust radiation from giant elliptical galaxies in their forma-
tion stage. The cosmological implications of such observa-
tions are obvious. We mayaiso envisage a galactic survey
for both the quasithermal emission from cool dust and tor
some higher rotational (and therefore optically thin)
transitions of the CO molecule. (But of course such
a survey would be carried out not with the VLT but with
small telescopes.) This should lead to a much more
realistic picture of the distribution of interstellar gas than
the present one which is based on the observation of the
opaque J = 1-+0 transition of 12CO.
The Space Telescope
The workshop was opened by three speakers-Longair,
O'Del1 and Macchetto-presenting a general introduction
of the technical as weil as the political aspects of the ST.
In his contribution "The Space Telescope and its
capabilites", Longair compared ST with ground-based
telescopes, stressing the improvements in angular resolu-
~
SCIENCE
INSTRlJ-lEl:TS ~ ••
-------
\ ,
I
J
I
I
/
/
 Space Telescope Parameters

Specification and performance goals of the NASA Space Faint-Object Spectrograph

Telescope
Teleseope aperture
System foeal ratio
Optieal design
Central obseuration
Field of view for seienee
Speetral range
Optieal performance:
system wavefront error
Radius of 70 % encireled
energy from a point souree
at 633 nm
Faint objeet sensitivity sub-
jeet to viewing in direetions
greater than
i) 50° from Sun
ii) 70° from Earth's limb
iii) 15° from Moon
Pointing stability
Minimum refleetivity
Speetral range 114 to 1010 nm
2.4 m Speetral resolving power R = 103 first priority
f/24 R = )"/6),, R = 10' seeond priority
Ritehey-Chretien Cassegrain System Absolute Effieieney > 1%
14 % aperture area Radiometrie preeision < 1 % of maximum scene signal
18 aremin diameter Siot sizes: fixed apertures 1.00 arcsec diameter
120 nm-l mm 0.50 arcsec diameter
0.25 arcsec diameter (fail-safe position)
),,/13.5 r.m.s. at 632.8 nm 0.10 arcsec diameter
Special purpose oeeulting mask
0.1 arcsec Closed
Centering stability ± 0.03 arcsec
Angular resolution 0.1 arcsec
Point objeets having m, = 27 or brighter;
extended sources to a limit of 23 magni- Dynamie range 10'
tudes (arsec) 'in visual wave-band. Both in Relative wavelength eali- 20 % of resolving power
10 hours of integration. bration
Deteetors 512 diode linear array Digieons
Exposure times Minimum gate time 50 Ils
0.007 arcsec on a target having m, = 13 in Minimum periodicity 10 ms
the range 400-800 nm
No limit on maximum length of target
85 % at 632.8 nm integration
70% at 120 nm

Satellite orbit
Cireular Earth orbit servieeable by the Spaee Shuttle
Altitude about 500 km
Inelination 28.8°
High-Resolution Spectrograph
Speetral resolvi ng power 2 x 10' 1.2 x 10'
The Wide-Field/Planetary Camera R = ),,/6),

The Camera ean operate with two different foeal ratios, 1112.9 or f/30. The Deteetor ehannel width:
former gives a fjeld of view 2.67 x 2.67 (aremin)' and is referred to as the at 150 nm 0.0075 nm 0.0013 nm
Wide-Field Camera; lhe latter gives a field ofview of 68.7 x 68.7 (arsee)'and at 230 nm 0.0115 nm 0.0019 nm
is ealled the Planetary Camera. Deteetor step or sub-step variable 0.001 variable 0.002
Wide-Field Camera Planetary Camera size - 2.3 nm - 0.38 nm
Pieture format 1600 x 1600 pixels 1600 x 1600 pixels Overall speetral range 105-170 nm 110-170 nm
110-320 nm 170-320 nm
Angular field of view 2.67 x 2.67 68.7 x 68.7
(aremin)' (arcsec)' Speetral range per integra-
Pixel size 0.1 arcsec 0.043 arcsec tion
at 150 nm 3.5 nm 0.56 nm
Dynamie range (per pixel. > 15.000 > 15,000
1.02 nm
at 230 nm 5.9 nm
single exposure)
Minimum and maximum in- 0.05 to 00 0.05 to 00
Maximum SIN (per pixel. 450 450
tegration times (seeonds)
single exposure)
Field view 1.0 x 1.0 (arcsec)' 1.0 x 1.0 (arcsec)'
Photometrie preeision Better than 1 % Better than 1 %
0.30 arcsec 0.30 arcsec
Photometrie aeeuraey Better than 2 % Better than 2 %
Dynamie range per resolu- 10' 10'
Overall dynamie range for + 3 to + 29 + 1 to + 29 tion element per exposure
stars (m,)
Photometrie resolution eount-limited eount-limited

Faint-Object Camera
This table shows the proposed instrumental parameters whieh will be
aehieved by the Faint-Objeet Camera operating in the f/96 mode.
Field of view
Angular resolution
Wavelength range
Dynamie range (single ob-
servation)
Photometrie aeeuraey
Foeal ratio
Number of pixels
Exposure time
Calibration
11 x 11 (arcsec)'
Essentially limited by the performance of
High-Speed Photometer/Polarimeter
the Optieal-Teleseope Assembly al wave-
lengths longer than 300 nm Speetral range 115-850 nm (photometry)
210-700 nm (polarimetry)
120 to 800 nm
Speetral resolution 2 nm (200-300 nm)
m, = 21 to 28 on point sourees; 15 10 22
- 25 nm (115-350, u, b, v, y, Hß)
visual magnitudes (are sec) , for extended
- 80 nm (UBVR)
sources
- 200 nm (110-350 nm)
- 1 % when not photon-noise limited - 500 nm (300-800 nm)
f/96 assuming 25 11m pixels. Additional fo- Fields of view 0.7, 1.4, 2.8 arcsec
eal ratios are desirable (see text)
Temporal resolution As fast as 16 I'S
500 x 500
Photometrie Pulse eounting over dynamie range of 10'
Up to 10 hours eharaeteristies with no dead-time eorreetion over first live
Internal sources and standard stars, neees- deeades; analogue mode extends range to
sary to provide required photometrie 10' and overlaps eounting mode over four
aeeuraey deeades with 2 % A/D eonversion aeeuraey.

19
tion, the "24-hours per day" observing time, and the wide
observable waveband. He described in some detail the five
scientific instruments: the Wide-Field/Planetary Camera,
the Faint-Object Spectrograph, the Faint-Object Camera,
the High-Resolution Spectrograph and the High-Speed
Photometer/Polarimeter. Longair concluded: "It is essen-
tial to plan your Space Telescope observing programme
now because there will not be time available for making
mistakes. Everything possible should be done from
ground-based observatories so that when the Space
Telescope observations are made, we are all in a postition
to make the optimum use of them."
O'Oell (NASA) and Macchetto (ESA) presented the status
of the project and showed that ST is not a telescope under
study but an instrument under construction. If everything
goes weil, we shall have a 2.4 m telescope in orbit in 1984.
So the rest of the conference was concerned with how to
make the best possible scientific use of it. For the benefit of
all interested parties we have reproduced on page 19 the
main parameters of the Space Telescope and its auxiliary
instrumentation.
Galactic Research
Oifferent aspects of star formation and stellar evolution
were the subject of interesting contributions by Appenzel-
ler, Gahm, Staller, Hack, Buser, Rakos and Jaschek. New
spectroscopic data obtained with IUE, a 45 cm space
telescope-rather primitive when compared with the
ST-were used to demonstrate the great potential of ST.
The unique possibility offered by ST to obtain extremely
deep pictures of star fields in different colours from the UV
to the IR will be used to investigate the space distribution,
luminosity function and chemical gradients of stars inside
and around our galaxy.
The H 11 regions, interesting because of their relation
with the star formation problem, opened the topics of
interstellar matter. Galactic as weil as extragalactic H 1I
regions, planetary nebulae and supernova remnants were
discussed by Courtes, N. Vidal, Elsässer, Perinotto, O'Odo-
rico and Oanziger. Making use of the advantage of the
monochromatic character of the emission of these objects
there was general consent that high-resolution pictures,
obtained with narrow interference filters, centred on lines
and on the continuum, will show new structures that could
revolutionize our ideas concerning the interstellar matter.
The main limitations will be in the angular sizes of the ST
fields of view, which are indeed very small when compared
with the extension of H 11 regions or supernova remnants.
Globular clusters are traditionally objects of transition
between galactic and extragalactic astronomy, and during
the ST Workshop they continued to play this role. The great
advantage of the use of ST in globular cluster research was
strongly expressed by Castellani. He concluded: "It looks
clear to me that ST could be devoted to such a problematic
for a very long time without exhausting this subject. I, of
course, cannot ask for a GCST (Globular Cluster Space
Telescope) though I am personally convinced that such an
instrument would be among the most busy and useful
instruments to be launched in space!". This idea was
immediately accepted by Auriere, Renzini and Wyller who
presented observing programmes for the ST for globular
clusters. Following these speakers, Weiss discussed the
use of globular clusters as distance indicator for galaxies.
20
Fig. 2: One of t"he main instruments onboard the Space Telescope
is the Faint-Object Gamera (FOG). It constitutes one of the major
European contributions to the ST project and is here shown (as
a full-size mock-up) at the European Space Agency (ESA).
Extragalactic Research
Westerlund pointed out that "many of the research
programmes that one may foresee for the Magellanic
Clouds with the aid ofthe ST are of a similar nature to those
that may be planned for objects in the Galaxy". Studies of
field stars, star associations, interstellar matter will make
significant contributions to the solution of the riddles of
the Magellanic Clouds. With ST it will be feasible to extend
similar programmes to other galaxies in the Local Group
and to the Virgo Cluster. PageI presented a review paper on
the chemical evolution of galaxies and van der Kruit
described the advantage of ST in the study of dynamics of
galaxies. In this field of astronomy one of the problems is
the velocity dispersion in elliptical galaxies. Crane de-
scribed the potential of the Faint-Object Camera in the
spectrographic mode with respect to dynamics of the
nuclear regions in elliptical galaxies.
If normal galaxies will be primary targets of ST, then the
so-ca lied "active" galaxies, including quasi-stellar and BL
Lacertae objects, are the extragalactic objects from which
we hope to obtain the most surprising results. Since long
one of the dreams of extragalactic astronomers is to
observe NGC 4157, 3C273 or OJ 287 with high spatial
resolution over a wide spectral range. ST will for the first
time allow us to look into the inner part of active nuclei of
galaxies close to "the sou rce".
These subjects for ST research were discussed exten-
sively by Ulrich, Penston, Oisney and Heidmann.
Tarenghi, N. Vidal and Fong dealt with distant galaxies
and clusters of galaxies and the puzzling problem of galaxy
evolution. The workshop ended with a review byTammann
on "Cosmology with the Space Telescope". In the begin-
ning, ST was considered "the tool" to solve all cosmologi-
cal problems. However, after a lucid analysis of the limits of
ST, Tammann concluded: "ST will be helpful for the
determination of Ho, it will be very important for the
mapping of the expansion field, and it offers a unique
chance to determine qo from supernovae and to verify the
Ooppler nature of cosmological redshifts."
The general impression after this workshop is that ST
will be a heavily oversubscribed telescope and that the
European astronomical community will try hard to obtain
more time than the 15 % share they are guaranteed in the
ESA-NASA agreement. M. Tarenghi
HO 101947-the First Very-Long-Period Classical Cepheid in
Our Galaxy?
W. Eichendorf and B. Reipurth

on La Silla, photometrically and spectroscopically with the

The ESO observatory on La Silla is not just a place
Strämgren photometer and the Bochum spectrum scan-
where you go to observe-it is also an important ner. Part of the observations were kindly performed by our
meeting place for European astronomers. Many successors after we left La Silla. As one result, we present
papers with co-authors from different institutes
have resulted from encounters over a cup of tea
and a "completo ". This was also the beginning of
the long collaboration between Drs. Walter Ei-
chendorf from the Astronomical Institute in 80-
07
chum, FRG, and 80 Reipurth from the Copenha-
gen Observatory, Denmark. Over aperiod of """'-,- ,,' .y lIu

several months, they continued to observe a star

that later turned out to be unique among the " .- .
Cepheids in our galaxy.
09 '.

The Oanish, German and Swiss telescopes on La Silla are

often used by "national" observers spending several 0.6
t:N
months on the mountain-in contrast to most other ESO '.

"',,,.,1 .
observers. This means that a huge observational material
can be secured for a programme, and also that such an "
observer at times can be recognized on the black circles
around his eyes at the end of an observing run. But it also
means that different types of programmes can be initiated, " 06
for example the study of very long-periodic phenomenae.
Meeting on La Silla and recognizing this fact, we .. ..
incorporated each night, in our respective photometric '. .
and spectroscopic programmes, observations of the
bright, 5m yellow supergiant HO 101947, which Fernie
(lBVS No. 1305) from observations on a few nights suspect-
ed to be a Cepheid with the long period of one month. Our
observations have shown this to be a conservative esti-
mate: We now believe HO 101947 is an extremely long-pe-
riod low-amplitude classical Cepheid with aperiod of ~06
lIb
'.
125 days, thus making it by far the longest-period Cepheid
hitherto known in our galaxy. In addition, this star may be
especially important for Cepheid research, because of its
membership in the small young open cluster Stock 14,
which has been observed by two Bochum observers, A. F. J. ". '. ,
Moffat and N. Vogt, in thei r study of southern open star
10
clusters.
All c1assical Cepheids known today in our galaxy have
periods ranging from 2 to 40 days, most of them falling in
the period range between 4 and 8 days. In the Magellanic
Clouds, however, Cepheids have been found with periods
up to hundreds of days. A few searches for such Cepheids 09
have been carried out in our galaxy-with no success so
far. Of course, heavy selection effects work against the .
discovery of these stars in our galaxy, in particular, if their
amplitudes are smalI. Recently several investigations
,
based on Cepheids have been published on possible
systematic differences between galaxies. Such differ-
1.1
ences, which might be due to chemical differences, could
even influence the calibration of the extragalactic distance I 1 I
scale. .JO 3550 3600 3650 3700
Observations on La Silla Fig. 1: The uvby lightcurve of HO 101947, which after the work
Between February and July 1978, HO 101947 was observed described here has been named V 810 Centauri. The diagram
with the Oanish 50 cm and the Bochum 61 cm telescopes covers half a year from January to July 1978.

21
here the differential uvby lightcurves (HO 101947 - HO
102350 in the instrumental system). Figure 1 covers half
a year. The internal accuracy of all measurements is better
than 0':'005.
We have observed two minima, and because,of the very
smooth behaviour of the light variations we can fairly safely
exclude that another minimum happened during the gap in
the observations. Therefore the period of the light varia-
tions of HO 101947 appears to be about 125 days, with
maximum amplitudes ranging from 0':'13 in y to 0~24 in v.
Ouring the first descending branch and minimum of the
lightcurve we obtained blue and red scanner spectra as
weil as H CL line profiles. On all spectra a clear H a emission
in the absorption core is seen, shifted redwards about
1.5 A. A careful check did not reveal any variations in the
spectra.

The Nature of HO 101947

It will obviously take same time to find out whether the
lightcurve is really periodic, and we shall here assume that
it repeats, a suggestion that may find some support in the
very smooth light changes. Without going into details here,
we found it very improbable that HO 101947 is a binary with
ellipsoidal variations, a SRd star or a RV Tauri star. The
spectral type, the luminosity class, an increasing blueness
with increasing brightness and the H a emission at least at
certain phases are all characteristics of long-period
Cepheids. Also the low galactic latitude of b = -0°38' is
typical for these stars. Aperiod of 125 days gives an
absolute magnitude of -8~3 in Sandage and Tammann's
period-Iuminosity relation, in good agreement with the
value of - r9 we get fram the distance of Stock 14.
Using a theoretical HR diagram from Cox and Hodson
(lAU Symposium No. 80) we find HO 101947 situated on the
blue edge of the Cepheid instability strip with nearly
perfect agreement between the theoretically expected and
the observed period.
Two features of the lightcurve cannot be readily ex-
plained: the small amplitudes, and the changing amplitu-
des. Concerning the first point we have already noted that
perhaps different chemical compositions would give Ma-
gellanic Cloud long-period Cepheids large amplitudes,
and galactic ones small amplitudes. Or, being situated on
the blue edge of the instability strip, HO 101947 may be
a first overtone pulsator, for wh ich smaller amplitudes are
theoretically expected, and it may even be switching to the
fundamental mode. Also, it has recently been found that
resonance phenomena may be effective for long-period
Cepheids. For example, the small amplitudes might result
from a coupling of a damped first overtone pulsation with
an excited fundamental mode. The problem of changing
amplitudes could perhaps also be understood as double-
mode pulsations, which are known from short-period
Cepheids.
The discovery of more stars in our galaxy behaving like
HO 101947 may help to salve this question. For other
galaxies, finding stars with amplitudes of only 0~2 over
months may be rather difficult at the moment.
Stock 14 and le 2944
The young, loose, open cluster Stock 14 is situated at
a distance of approximately 2.5 kpc on the innersideofthe
Carina spiral feature. Only 1 degree away is the cluster and
large H 11 region IC 2944. At the IV. European lAU meeting
in Uppsala A. Ardeberg and E. Maurice reported that
IC 2944 actually consists of at least seven aggregates
Fig. 2: An equidensity-picture of the region around IC 2944 and
Stock 14 in red light. North is on top. In the lower left the large H 11
region IC 2944 is seen; the bright star indicated by an arrow is
HO 101947. It is evident that IC 2944 extends at least to a very small
angular separation from Stock 14.
This picture is an overlay of two equidensity-pictures (one of
first and one of second order) of a red plate taken by G. Lyng!l
(Lund). In this way the H 11 region is very dark, but the fine
extensions towards Stock 14 are easier to see than in a usual
photograph. For the picture we are grateful to the staft photo-
grapher in Bochum, Mr. W. Hünecke.
stretching out along the line of sight. The one closest in
angular distance to Stock 14 also seems to be at compar-
able distance, suggesting that they may be physically
connected. We have made a preliminary search for faint
bridges of matter between Stock 14 and the agg regates in
IC 2944 using the technique of equidensitometry on deep
red plates, taken by G. Lyngä. Figure 2 shows an example
of the pictures obtained that way, and it is seen that part of
the H 11 region in IC 2944 is stretching out at least to small
angular distances from Stock 14. A more detailed study
with a microdensitometer on more plate material is in
preparation. Also radial-velocity studies in the region will
be important to salve this question.
Ta confirm the Cepheid nature of HO 101947, more
observations over lang-time intervals will be necessary,
both photometrical and spectroscopical. For a precise
understanding of the pulsations, observations over more
than a decade may perhaps be needed. We shall certainly
try to continue observing HO 101947, but we here want to
take the opportunity to urge observers in the southern
hemisphere to include this exciting star in their photo-
metric and spectroscopic observing programmes.
Financial support by the German (OFG, Oa 75/3 and
Schm 160/13) and Oanish research council is gratefully
acknowledged.

22
Sulfur Dioxide and Carbon Disulfide in the Venus Clouds
C. T Hua, G. Courtes and N. H. Doan
Exciting results were obtained trom the Pioneer
Venus spacecraft as it reached our mysterious
neighbour planet in late 1978. In contrast to our
own lite-supporting, oxygen-rich atmosphere,
that on Venus is dense, hot and poisonous.
Spectroscopic observations near the atmospheric
cut-ott at 3000 A were recently made at La Silla by
Drs. C. T. Hua and G. Courtes (Laboratoire d'As-
tronomie Spatiale, Marseille), together with Dr.
N. H. Doan (Observatoire de Lyon), showing tor
the tirst time that there may be carbon disultide in
the Venus clouds, adding a compelling malodor-
ous reason tor not going therel
We have recently obtained very interesting spectroscopic
data concerning the clouds on Venus, in the range of
3100-3300 A and with aresolution better than 2.5 A/mm.
The observations were carried out by means of a scanner
attached at the Cassegrain focus of the ESO 1.52 m
telescope, on January 27,1979.
~ \
\
t;)._-~
(J)
C
~ Iy by observations carried out near the horizon, at middle
-t-'
C
a
X X X Xx
X
XX x
3100 3190
The ultraviolet spectrum of Venus, as observed with the ESO 1.52 m telescope. Tracing (b) shows the intensity ratio (Venuslsky
background) and the absorption bands in the Venus atmosphere. The crosses (a) indicate the ratio between two separate sky
measurements (see text) and that the effects of the ozone absorption bands in the terrestrial atmosphere have been cancelled.
A detailed examination of the spectra seems to indicate
the presence of carbon disulfide (CS 2), rather than sulfur
dioxide (S02) as previously thought; this conclusion is
based on an extensive investigation of the UV absorption
of S02 and CS 2·
A preliminary study concerning the optimization of the
optical arrangement for the VENE RA experiment of the
Service d'Aeronomie and CNES proposed by Prof. J. E.
Blamont facilitated our choice of the best observational
parameters for our present investigation of the Venus
clouds. In particular, it was found that the contrast in the
UV spectral range near 3300 A appears to pass through
a maximum; this was deduced by examining various
filtered photographs obtained at the Pic-du-Midi Observa-
tory. However, until recently few ground-based or space
experiments have been carried out with the aim of
explaining the reason for this increase in contrast, i. e. why
the clouds are better visible in this spectral range, i. e. near
the atmospheric cut-off at 3000 A.
Inorder to identify the natu re of the absorbing molecular
bands which are the probable cause of this contrast
phenomenon, the best method is evidently to analyse
high-resolution spectra. This is because direct photo-
graphs from the ground are limited by the atmospheric
turbulence, the rapid variation of atmospheric extinction
and Venus's low ultra-violet intensity. Moreover, it is
difficult to have simultaneous good conditions for obser-
vations of Venus and the photometrie standard star. The
spectrophotometric method is therefore the most conve-
nient and suitable.
One of the likely candidate moleeules for the Venus
clouds is sulfur dioxide, since it has been reported by
several authors that aqueous H2S0 4 solutions of high
concentration are required to explain the stratospheric
dryness of the planet. The presence of this acid aerosol is
deduced from polarimetrie and spectroscopic data, espe-
cially in the IR spectral range. Furthermore, the decrease of
the ultraviolet reflectivity of Venus has been interpreted as
being due to the presence of particles of elemental sulfur
inside the clouds (Hapke and Nelson, 1975).
Unfortunately, in ground-based observations, the S02
ultraviolet spectrum coincides quite weil with the absorp-
tion banks of atmospheric ozone between 3000 A and
3300 A. This makes its detection rather difficult, particular-
or high geographie latitudes and at low altitude sites. In
addition, this nearly perfeet wavelength coincidence re-
sults in a modulation enhancement of the ultraviolet
32BO A 3400 "-
23
spectrum of the Venus clouds, the intensity of wh ich is
already very weak at the limit of transparency of the
terrestrial atmosphere.
However, a previous study by Barker et al. (1975)
reported the presence of a broad absorption feature
between 3200 A and 3100 A, from relatively low resolution
spectra (10 A). But no identification of the c10uds was
made. Recently, Young (1979), discussing these above
data, suggested thatthe 3150 A band could be attributed to
carbon disulfide (CS 2)·
The Observations
We observed the ultraviolet spectrum of the Venus clouds
by means of a scanner with a rather higher resolution
(2.5 A). Our recent report (Hua et al., 1979) also described
a method allowing us to derive the intrinsic spectrum of the
planet despite the strong terrestrial ozone absorption and
thus to measure the absorption bands, presumed attribut-
able to S02' At the same time, Stewart et al. (1979), in an
article concerning preliminary results from the NASA
Pioneer Venus Orbiter, found two broad absorption fea-
tures near 2100 and 2800 A, fitting weil the known S02
absorption bands. The main advantage of our study,
however, is the use of a scanner with a higher resolution
that is capable of resolving the individual molecular bands
(spaced at about 20 A interval), as compared to the 13 A
resolution of the Pioneer UV spectroscopy.
Tracing (b) in figure 1 shows the ratio I), (Venus)/I, (sky
background) plotted against wavelength, as obtained on
January 27,1979, with a photoelectric scanner attached at
the Cassegrain focus of the ESO 1.52 m telescope. Since
the atmospheric ozone absorption (the origin of the
atmospheric cut-off at 3000 A) influences the sky back-
NEWS and NOTES
Spectra from the ESO Schmidt Telescape
One 01 the most efficient means for astronomical spectroscopy is
the objective-prism method, cl. the work at the GPO astrograph
described in this issue of the Messenger, pages 10 and 29.
One 01 the largest objective-prisms in the world is the 1 m
ultraviolet-transparent prism at the ESO Schmidttelescope. This
prism gives a dispersion of about 450 Ä/mm near Hy at 4340 Ä. It
was provided lor stellar spectroscopy and has already proved its
great value in connection with extensive classification pro-
grammes.
With a moderately high dispersion in the blue and violet, the
ESO prism still has a reasonable ~ispersion in the red, about
1500 A/mm between 5000 and 7000 A. It is therelore an extremely
uselul device lor search programmes that aim at linding objects
with f. ins!. Ha-emission at 6563 Ä and it has already yielded many
new planetary nebulae and other emission objects in the Milky
Way.
Now, however, it appears that there is an even more important
field for the ESO Schmidt telescope. For several years now,
quasars have been lound with the smaller Gurtis Schmidt
telescope at Gerro Tololo, just south 01 La Silla, and in Australia
with the SRG 48" Schmidt telescope at Siding Spring. These
quasar searches were carried out by means 01 III a-J plates, which
are sensitive to blue li~ht and with prisms 01 relatively low
dispersion (1800 to 2400 A/mm near Hy). For this reason, all 01 the
quasars that have been lound have redshilts less than about 3.5,
corresponding to the Lyman 0 line at1216 Ä redshifted toaboutthe
red limit of the 111 a-J emulsion, near 5500 Ä.
24
ground spectrum as weil as that of Venus, the intensity
ratio of two sky background spectra, one recorded near
Venus and the other near the western horizon at the same
zenith distance, should be the same for all wavelengths, cf.
tracing (a). The small residuals we observe can be ascribed
to local differences in the thickness of the ozone layer.
Hence, the absorption features that are seen in the
Venus/sky ratio (b), are only attributable to the Venus
clouds. Two of the absorption bands are centred at 3150
and 3170 A and coincide with known S02 bands. But the
strongest S02 bands are observed at somewhat shorter
wavelengths and it therefore appears that it is more
reasonable to connect the observed features at 3150, 3190,
3204, 3235 and 3275 A to CS 2, involving the vibrational
structure. The electronic absorption spectrum of CS 2 has
been extensively investigated in the past, cf. the V-system
in the 2900-3300 A range.
We presently need more observations to confirm the
above preliminary conclusion. It would be worth while to
extend this kind of analysis to longer wavelengths. We are
now examining several 3 A/mm coude spectra from the
1.52 m telescope of good quality (taken by P. Bouchet and
N. Bahamondes) and we hope soon to be able to present
further details.
References
Sarker, E. S., Woodman, J. H., Perry, M.A., Hapke, S., Nelson, R.J.,
1975, J. Atmos. Sei., 32, 1205.
Hapke, S., Nelson, R., 1975, J. Atmos. Sei., 32,1212.
Hua, G. T. Gourtes, G., Doan, N. H., 1979, C. R. Aead. Sei., Paris,
B 46, in press (19 March).
Stewart, A.I., Anderson, Jr., D. E., Esposito, L. W., Sarth, G.A.,
1979, Seienee, 203, 777.
Young, A. T., 1979, learus, 37, 297.
Searches for quasars with higher redshifts have not been
possible with these telescopes, because 01 the very low dispersion
01 their prisms in the red. Sut, of course, the results in the blue
have already had a profound impact on our knowledge about
quasars, and more than hall 01 the known quasars have been
lound with them, including some of the most intriguing ones with
absorption lines, etc. However, the 1 m Schmidt telescope at La
Silla has a sullicient dispersion (or rather resolution) in the red to
make a high-redshift quasar search possible. In addition to the
dispersion, the large field (5.5 x 5.5 degrees) and the faint limiting
magnitude are important parameters because nobody expeets to
lind many high-z quasars (if any at all). Remembering that the
redshift, accordi ng to the majority of astronomers, is a measu re of
distance, it therefore seems as if the ESO Schmidt has a unique
potential for looking further out into the vast expanses of the
Universe than most other telescopes.
The first very deep plates have now been obtained. 180-minute
exposures on red-sensitive 111 a-F emulsion show a wealth 01
objects 01 diflerent spectral classes and are now being searched
by a number of astronomers, in ESO and outside. No high-z
quasars have been found so far, but the search has just begun. It is
probable that any Lyman a-emission lines near6000 Ä or further to
the red will be very broad and maybe also rather shallow.
Unfortunately, the 111 a-F emulsion is not equally sensitive to all
wavelengths, there are spectral regions where it is particularly
sensitive and which therefore appear as "humps" on the other-
wise "flat" spectra. These features can easily be confounded with
emission features, and careful measurements are necessary to
confirm whether they are intrinsic, i.e. belong to the astronomical
A double emission-fine galaxy recorded on a 180-min exposure with the ESO 1 m telescope, on /IIa-F emulsion behind a GG 385 filter.
The spectral region is roughly 3800-7000.4. Several emission lines are seen, i. e. Ha (6563.4), the {O /IIJ doublet at 5007 and 4959 Ä, the
Hß line at 4861 .4 and, at the ultraviolet edge, (O IIJ (3727.4). These spectra are unwidened in order to reach the faintest possible limiting
magnitude. Observer: H.-E. Schuster.

object, or whether they are just artifacts of the 111 a-F emulsion.
In any case, a candidate object must first be studied with a slit
spectragraph at much higher dispersion before it can be said to be
a quasar.
Let us ho pe that it shall soon be possible to report some further
news about this important research. Meanwhile, as an illustration
of the possibilities offered by this technique, we show a small part
of one of the I1I a-F prism plates with a double galaxy that clearly
has a very strang emission-line spectrum. Many more are seen on
these plates.

PERSONNEL MOVEMENTS
(A) Statt Heino WIRING (German). Internal Auditor, 30.6.1979.

ARRIVALS Roman MARCINOWSKI (Belgian), Accountant, 31.8.1979.

Geneva Geneva
Bernard GUTTIN-LOMBARD (French), Laboratory Technician
Sölve ANDERSSON (Swedish), Electronics Technician, 24. 4.
(Electronics), 31.7.1979.
1979.
La Silla La Silla

Bernard BUZZONI (French), Optical Technician, 1. 4.1979. Max Jean L1Z0T (French), Optical Engineer, 31.7.1979.

Ray WILSON (British), Head of Instrument Development Group; Fred SUTER (Swiss). Systems Analyst/Pragrammer, 31.7.1979.
transferred from Geneva, 15. 6. 1979. Sten RÖNNBOM (Swedish), Electranics Technician, 31. 8.1979.
Egon HESSENMÜllER (German), Optical Technician,
1.7.1979.

DEPARTURES (8) Temporary

Garehing La Silla
Ursula LIESE (German), Shorthand-typist (telephone and telex Christopher SMITH (Canadian), Resident Astronomer, from
operations), 31. 3.1979. 16.5.-31.8.1979.

25
First Photographs of Andromeda and Galactic Clusters at 1950 A
M. Golay, D. Huguenin, Geneva Observatory; M. Deharveng, LAS, Marseille

••• .,
~)'
, '"" .
.."'"
• 0. "". \

" .
,
.. :.
Fig. 1: 10-minute exposure of Andromeda (M31) at 1950.4 (Ieft). To the right a visual photo (5500.4) for comparison.

The photographs shown in this note were taken on Oetober 30, The eamera is eomposed of a Sehmidt-Cassegrain objeetive, an
1978, with a eamera built by the Spaee Astronomy Laboratory of ultraviolet image intensifier and a 35 mm film holder. Here are
Marseille (LAS) and launehed in the stratospherie gondola of several other eharaeteristies of the eamera and the gondola:
Geneva Observatory.
Camera:
This was the seeond flight of a long series of launehings the
Diameter of the aperture: 130 mm
main purpose of whieh is photography at 1950 'A of the entire
galaelie plane and eertain regions near the northern and southern Foeal length: 230 mm
galaetie poles. More than 30 photographs are presently being Equivalent photometrie
analysed. B stars down to 13th magnitude are observable, and aperture: F: 2.10
some galaxies have already been deteeted in the ultraviolet, as
Field: 6° diameter
weil as a number of blue, galaetie halo stars.
The three pietures presented here are of partieular interest Passband: 1900 'A-2075 'A obtained by selee-
beeause they are the first photographs in the ultraviolet of a spiral tive eoating of 2 mirrors. Bloeking
galaxy, of a pair of young galaetie clusters and of a supernova in the visible by the es-Te eathod
remnant whieh exploded 50,000 years ago. Apart of this of the image eonverter
nebulosity is also an X-ray source. For the first time a fine strueture
of a spiral galaxy ean be analysed in the ultraviolet. The nueleus of
Andromeda (M 31) in the ultraviolet is mueh smaller than on the
plate taken in the visible. The intensity of the nueleus ean be
esti mated to be about 5 x 10 "erg em 2 sec I 'A I at 1950 'A.

Fig.3: The young double cluster, h andXPer at 1950 Ä. Exposure

Fig.2: The Cygnus Loop. Exposure time 1 minute. time 5 minutes.

26
Film: Kodak
Construction of the ESO
Weight: 17 kg
Headquarters Building
Gondola:
The Observatory of Geneva system eonstrueted in eooperation
with the Spaee Astronomy Laboratory of Marseille. The Max-Planck Society informed ESO that the construc-
tion company in charge of the ESO project has unexpect-
Guiding: single axis by servo-guided side-
edly run into financial difficulties and that a new con-
rost at of 300 x 300 mm
struction fi rm has now taken over the project.
RMS preeision of 20 are seeonds
magnitude range m B = - 2 to + 6.4
There will be some delay but the termination of the
Total weight: 337 kg building is still expected for summer next year. The exact
Flight altitude: 40 km, balloon of 350,000 m' date is 31 July 1980.

Cassegrain Echelle Spectrograph (CASPEC)

M. /e Luyer, J. Me/nick, W. Richter

2. Optical Concept
The CASPEC figures prominently among the The concept is based on a 15 cm echelle grating and
future, highly advanced auxiliary instruments for a plane cross-dispersion grating which provide two-di-
the ESO 3.6 m telescope. It will allow high-disper- mensional spectra.
sion, spectroscopic observations of comparative- The instrument has been designed to be used in three
Iy faint objects to be made in a reasonab/e amount different modes (resolving powers 17500, 30300 and
60600) as shown in table 1. Shown also is the required
of observing time. When it enters into operation in combination of echelle grating, cross-disperser grating
late 1980, it will become possible to analyse and camera for each of these modes.
distant stars and nebulae in great detail. The The principal detector will be a SEC vidicon tube, which
CASPEC project is directed by Maurice le Luyer, is described on page 34. This tube has a target area of 25 x
Jorge Melnick and Wolfgang Richter from ESO 25 mm and a pixel size of 25 flm. The last lines of table 1
give slit width and length (and the corresponding angular
Geneva. resolution on the sky) per pixel. The optical scheme uses
the minimum number of elements to get a very efficient
white instrument. It comprises 3 mirrors, 21enses and
CASPEC is the fi rst major instrument for observation at the
Cassegrain focus of the 3.6 m telescope which has been
designed by ESO and which is now going into manufac- Table 1. Optieal Parameters
ture. This seems to be the appropriate moment to describe
the main features of the instrument and to hope for Resolving
17500 30300 60600
power
eventual comments from the future users, comments
which are useful for the finalization of the instrument.
Dispersion
at = 5000 Ä 9.5 Ä/mm 5.5 Ä/mm 2.8 Ä/mm
1. Astronomical Purpose
Placed at the Cassegrain focus of the 3.6 m telescope, the Echelle grating Jobin Yvon Bausch and Lomb
CASPEC"will provide astronomers with spectrograms of blaze angle 46°30' 63°26'
resolutions previously obtained only with large coude line pairs 95 mm ' 79 mm ' 31.6 mm '
instruments, where a significant fraction of the light is lost
in additional reflections required to bring the beam to the Cross disperser
remote coude focus. The considerably higher dispersion blaze angle 4°18'
of echelle gratings as compared to conventional grating line pairs 300 mm '
spectrographs makes high-resolution work possible atthe
Cassegrain focus. The CASPEC will allow observers to White camera
obtain high-resolution spectra of objects much fainter foeus F = 279 mm 560 mm
than what would normally be possible with a coude aperture 1/1.66 1/3.3
spectrograph.
The possibility of obtaining high-resolution spectro- Resolution/pixel
scopic observations of stars as faint as 15 magnitude will slit width 144 f.l 173 fl 86 f.l
open a vast field of research to European astronomers, in skyangle 1" 1"2 0"6-
particular since very high resolution studies of the proper-
ties of galactic and nearby extragalactic stellar and slit length 192 fl 277 fl 138 f.l
interstellar systems only observable from the southern skyangle 1"3 1"9 1"
hemisphere will become possible for the first time.

27
 f) 1100

10

1·- 8

5-

_ _ _ 660"---- -1 395

Fig. 1: CASPEC Assembly: (1) slit area Fig. 2, (2) col/imator, (3) Hartmann mask, (4) echel/e gratings, (5) cross dispersergrating, (6) short
camera. (7) long camera, (8) detector. (9) comparison lights. (10) slit rear viewer.

2 gratings. The wavelength range is limited by the efficien- 4. User Interface

cy of the detector, 0.38-0.65 ~ for the chosen photo-
cathode of the vidicon.
3. Mechanical Concept
The basic concept is a straight (not folded) very compact
layout, leading to a fairly small and strong housing with
modules for the different functions as shown in figure 1.
To change from one mode of observation to another
requires the exchange of modules. Such a module is
a grating with cell, adjustment mechanism and base-
flange. Another module is a camera with detector and
base-flange.
The echelle grating has a manually operated adjustment
screw to centre the spectra on the detector. The tilt of the
cross-disperser grating which is used to centre the desired
wavelength band on the detector has remote contro!.
The slit area is one unit wh ich is composed of several
modules as shown in figure 2. This gives the possibility for
a later exchange of some modules by others. All modules
are equipped with remote control (motor and digital
read-out). The adjustment ranges are given in t,able 2.
Table 2. Adjustment Range
Function Range
Slit opening 0.07 2 mm
Dekker opening 0.1 20mm
Collimator focus ± 3.5 mm
Cross disperser ± 5°
The status of the instrument is completely controlled
through the Instrument Computer and the standard Casse-
grain area peripherals. The philosophy is very similar to
what is being made for the Coude Echelle Spectrometer.
A touch panel in front of a display plays the role of
a "software" buttons set-the function of each button is
given on the display.
A second display shows the status ofthe instrument. The
form-filling technique simplifies strongly the introduction
of new parameters.
A third display, which is a graphics one, permits the user
to "play" with the data.
5. Data Reduction
The advantage of using a square format has to be paid for
bya rather complex data reduction programme. A meeting
was organized on this subject last March. J. Melnick, the
chairman, reports an page 13 on the discussions of this
meeting.
6. Assistance from the Review Team
The Review Team members (J. Andersen, L. Delbouille,
E. Maurice, P. E. Nissen) received a very detailed informa-
tion manual and a complete set of sub-assembly drawings.
This gave rise to most fruitful comments and discussions,
especially in the following aspects: flexure analysis, con-
trols and electronic interface, data processing (one RT
member initialed a small ESO workshop on this subject),
image slicer exchange, Hartmann mask design. We feel we

28
have been very much guided by this excellent cooperation

I
ofthe RT members to whom we extend our grateful thanks. Light beam

To slit viewer
7. Present Status (2.5.79)
The design is near completion and the bids for optical and \ .
mechanical parts have been received. The contracts for
manufacture are in the process of being signed. Handle I \ I

8. Future Plans
After the definition of the test procedures we have now
-
-
Comparison
light
~
'/-f'~\:
I
11 I
11>.,
'--rr
11
I
I
.... '11
started with the design of the test facilities. One part 1
comprises the tests in Geneva (stability of spectrograph,
functional test of the detector and of the instrument), the Filter wh<><> I

-+E-·~~~~~d~\-E.ill.e.LW.becl
other for tests on La Si Ila. Also in the near future the control
and handling aspects will be finalized.
The present planning of the main activities for the
subsequent development is shown in table 3. pre slit Dekker r I - .L"\lj[·~ _.-}- Slil

Table 3. Present Sehedule .1-. Filtpr wheel

Aetivity Dates
-m-·--f I.
I
r
:'
Manufaeture meehanies
Assembly meehanies
May-Sepl.
Oet.
1979
1979 "'- ~ /17>
f----.--
"'-. ,>f/ . 1 Rear slil
Assembly eontrols Nov.-Dee. 1979 ~?< vlewer
Manufaeture opties May-Dee. 1979
Opties test Geneva Jan.-Feb. 1980
Improvements Mareh-May 1980 Foot I
Integral test June-July 1980
Shipment Aug.-Oel. 1980
Installation La Silla Nov.-Dee. 1980
To eollimator

Seale.1: 2,5
Fig. 2: 51it area. ~

Relative Radial Velocities of Stars Determined

from GPO Spectrograms
F. Gieseking

meters for the investigation of the kinematics and dy-

One of the first telescopes to be installed on La
Silla was the GPO 40 cm astrograph. Although it is
one of the "smallest" instruments at the ESO
observatory, it is by no means less productive than
the larger ones! On the contrary, the impressive
results that have recently been obtained by Dr.
Frank Gieseking of the Hoher List Observatory,
near Bonn, proves the tremendous potential of
this instrument. Thanks to the good accuracy of
the measured radial velocities, large-scale investi-
gations of stellar motions can now be carried out.
The General Problem
The radial velocities and the proper motions of stars (and
their possible temporal variations) are fundamental para-
namics of stars and stellar systems like binary and multiple
systems of stars, stellar associations, star clusters, the
galaxy as a whole and clusters of galaxies. Proper motions,
however, defined as the angular velocity of the tangential
component of the space velocities, are important only in
the solar neighbourhood. For example, a binary system,
even with the large orbital period of 100 years, consisting
of two solar-type stars, has aseparation of less than 0.1
arcsecond if at a distance larger than 270 parsecs. With
terrestrial telescopes such a system can only be resolved
by application of laborious interferometric techniques like
speckle interferometry. On the other hand, the error of the
tangential motion of si ng le stars as derived from proper
motions is proportional to their distances. Typical values
for the errors of average accurate proper motions indicate
that already at distances larger than 800 parsecs the error
of the tangential velocity is larger than ±6 km S-I.
Contrary to this, the error of radial-velocity measure-
ments is in principle, independent, of the distance of the

29
stars. Since the accuracy of radial-velocity determinations
from spectra of medium dispersion (60 A mm ') is of the
order of ± 6 km s ' (for early-type stars), we arrive at the
conclusion that at distances larger than 800 parsecs the
radial component of the space velocity can be determined
more precisely than the tangential component. (At still
larger distances the radial velocity soon becomes the only
relevant kinematical parameter of stars and stellar sys-
tems.)
Unfortunately, however, at distances larger than
800 parsecs even the intrinsically brighter stars (B 1 V) are
fainter than 8.5 mag in V. Since the measurement of radial
velocities with conventional slit spectrographs is very
laborious (requiring long observing times at large tele-
scopes for the exposure of only one spectrum at a time).
our data for stars fainter than 6 mag are still rather
incomplete. We therefore co me to the general conclusion
that a more efficient method for radial-velocity determina-
tions is urgently needed for a further progress in the
investigation of the kinematics and dynamics of the stars.
The Fehrenbach Prism
The search for more efficient methods to determine stellar
radial velocities is the search for more efficient methods to
obtain stellar spectra. The most efficient astronomical
spectrograph is the objective prism-telescope combina-
tion, because:
(1) The light losses are reduced to a minimum: In detail,
the light losses of an objective prism by reflection and
absorption are expected to be of the order of 30 %,
whereas in the case of the conventional grating slit
spectrograph light losses by reflection of about 60% and
transmission losses of about 85 % are typical (Fehrenbach,
eh.: 1966, Adv. Astron. Astrophys.4, 1). (For slit spectro-
graphs without image slicer the sometimes drastic light
losses at the slit must be added.) This means that an
objective prism has at least a tenfold greater transparency
as compared with a conventional slit spectrograph.
(2) The objective prism allows the exposure of nu-
merous stellar spectra at the same time: In detail, in
Fig. 1: An enlarged 1.2 by 0.46 square-degree section of plate GPO 2557, which is approximately only 1/10 of the original plate. At top is
the western part of the rich open star cluster NGC 3532. The total plate, exposed for 2 times 30 minutes, shows approximately 500 pairs
of spectra. The spectra of more than 100 stars down to the 10th magnitude are exposed to sufficient density to be measured with high
precision. The marked star is HO 95878 (see fig. 2). North is to the right, west is down.
30
a 4-square-degree field of a telescope of 4 meter focal
length, well-exposed 100 Amm 1 spectra of typically up to
100 stars can be expected.
From this example we see that an objective prism should
be a thousand times more efficient than a conventional slit
spectrograph. (For a realistic comparison for the case of
the GPO astrograph, see below.)
Unfortunately, there are two serious drawbacks of the
objective prism:
(1) Due to the absence of any comparison spectrallines,
the wavelengths of the objective-prism spectra are apriori
not calibrated.
(2) Since we have to do with a slitless spectrograph, the
spectral resolution depends on the seeing.
Whereas the second problem cannot be overcome by
terrestrial telescopes, there have been many attempts to
calibrate the wavelengths of objective-prism spectra. After
numerous failures, Pickering's (Astron. Nachr. 142, 105,
1896) "reversion method" proved to be most attractive. Its
principle is as folIows: First, an exposure of the star field is
made through the objective prism adjusted with the
direction of dispersion parallel to declination. Then, the
exposure of the same star field is repeated on the same
plate after "reversion" of the prism by rotating it through
1800 around the optical axis of the telescope. If the
telescope has been slightly shifted in right ascension
between the exposures, the plate contains two side-by-
side spectra of each star, with the red end of one next to the
violet end of the other. It should now in prineiple be
possible to measure Doppler shifts of the speetrallines by
eomparing the positions of lines in one speetrum with
those of the eorrespondi ng lines in the reversed speetrum.
But unfortunately, the relative positions of the lines in
both spectra are not determined by the Doppler shift alone,
beeause:
(1) They are partly determined by the performance of the
reversion (the deelinations of the plate eentres of the
normal and the reversed exposure eannot be made exaetly
the same [an aceuraey of typieally better than 0.05
arcseeond would be neeessary)) and by instrumental
influenees like temperature effeets, meehanieal stability,
ete.
 (2) The distorsion of anormal prism produces an
apparent shift of the spectral lines, which in a 2-degree
field is of the order of ± several thousand km 5 '!
The first problem, which cannot be overcome in prac-
tice, limits us to the determination of only relative radial
velocities. Because of the second problem, the application
of objective ·prisms for the determination of radial veloci-
ties was forgotten for several decades, until Charles
Fehrenbach invented the distorsion-free, direct vision
Fehrenbach prism. In simplest form, it is a compound
prism, shaped like a plane-parallel plate, which consists of
two prisms made of different kinds of glass. The compo-
nents have different dispersive powers, but equal refrac-
tive indices for a certain wavelength in an accessible
wavelength region. (For details the reader is referred to
Fehrenbach [Ann. d'Astrophys. 10, 257, 1947] or to the
popular representation by Gieseking [Sky & Tel. 57, 142,
1979]).
The radial-velocity astrog raph of the European Southern
Observatory is equipped with one of the six Fehrenbach
prisms existing today. This "Grand Prism Objectif" (GPO)
has a diameter of 40 cm and produces at 4 meter focal
length spectra with a linear reciprocal dispersion of 110 A
mm '. With this instrument, the reversion method can now
be applied successfully. The latest refinement of this
method was supplied by the author, who introduced
a modified reduction method for Fehrenbach-prisms spec-
trograms, which yields a considerable improvement of the
measuring accuracy (Astron. Astrophys. 47, 43, 1976 and
Sky & Tel. 57, 142, 1979).
Present Results
With this new reduction method, until now a total of about
8,000 relative radial velocities of more than 500 stars in
10 star fields (mostly in the region of selected open star
clusters) have been determined. The investigated stars
have photographic magnitudes fainter than 6 mag and
brighter than 10.5 mag. Their spectral types are mostly
between late Band early A. As expected, the measuring
accuracy depends on the density of the spectra and
therefore (with the exposure time as an additional parame-
ter) on the brightness of the stars. With an exposure time of
2 times 30 minutes the observation errors of the radial
velocities of stars brighter than 9.7 mag lie between 4 km
5 ' and 9 km 5 '. This is approximately half the error to be
expected for radial-velocity measurements of early-type
stars from slit spectra with a comparable dispersion of
110 A mm '. Since the strictly relativemeasuring principle
of the reversion method with a Fehrenbach prism is
affected mainly by statistical observation errors, the accu-
racy can be significantly improved by accumulation of
observations. This is contrary to the slit-spectrum tech-
nique, which is also affected by systematic errors. (This is
demonstrated by the fact that the external error of the
measurements is expected to be twice the internal error.)
Since for each star on the average 15 observations are
available, the observation error for about 300 stars with
(assumed) constant radial velocities can be brought down
to the region of ± 2 km s '. Herewith the quality of the
relative GPO velocities of stars fainterthan 6 mag becomes
comparable to that of most published "conventional"
radial-velocity data of stars brighter than 6 mag.
For a realistic estimate of the efficiency of the GPO
method we have the opportunity to compare for a number
of stars in NGC 6475 the GPO radial velocities with (by
chance) almost equally accurate slit-spectrum measure-
ments. Though the slit spectrograms were obtained with
a fast grating spectrograph at a 90 cm reflector, the GPO
was faster by a factor 5 per star, after reduction to equal
telescope aperture. After inspection of the whole GPO
material, on the average 35 stars per field were found,
which could be measured with the accuracy in question.
Therefore, the total gain in efficiency is of the order of 150
to 200 (which in rich fields could further increase by
a factor of 2).
We therefore come to the conclusion that the GPO
method offers the first really promising, efficient method
for the determination of the radial velocities of the fainter
stars.
The present GPO material has so far been analysed with
two kinds of study in mind: The first aspect is the
investigation of spectroscopic binaries. If the dependence
of the measuring error on the density of the spectra is
considered carefully, the GPO data establish the most
homogeneous, large set of relatively accurate radial
velocities published so far. Therefore, the identification of
variable radial velocities is possible with very high statisti-
cal significance. A careful statistical analysis of the
observed overall error distribution yielded the first reason-
able estimate of the frequency of spectroscopic binaries.
This number may provide an important contribution to the
complex problem of the overall binary frequency and the
frequency distribution of the semi-major axes of binaries,
which in turn is a measure of the frequency distribution of
angular momentum. Besides the possibility of a statistical
identification of spectroscopic binaries with even extreme-
Iy small amplitudes, an individual identification of spectro-
scopic binaries proved to be reliable down to radial-veloci-
ty amplitudes of K = 12 km s '. The observations of one of
the numerous spectroscopic binaries discovered in this
way on GPO plates are illustrated in the figures with this
article.
The second aspect is the investigation of the kinematics
and dynamics of open star clusters. So far only the study of
the most extensively observed cluster NGC 3532 (see fig. 1)
has been completed. Here, relative radial velocities of
a total of 82 field and cluster stars could be determined
with a mean error of only ± 1.77 km s '. With these GPO
measurements, NGC 3532 becomes the second open star
cluster (besides the Pleiades), for which accurate radial
velocities of a larger number of stars are known. Moreover,
NGC 3532 is the first open cluster for which reliable
membership probabilities were calculated on the basis of
a statistical analysis of the observed radial-velocity distri-
bution of the stars. Besides this essential contribution to
the separation of field and cluster stars-which is espe-
cially important in the halo of the cluster-the GPO radial
velocities proved to be accurate enough to estimate even
the true internal motion of the cluster, as far as the
investigated cluster stars are concerned. This, however, is
the key to the dynamical properties of the aggregate, like
its dynamical mass, the mass distribution or the structure
of the cluster halo and the evaporation of cluster stars.
These results will be published in aseries of forthcoming
papers.
Future Programmes
First of all the investigation of NGC 3532 should be
extended to fainter stars. Furthermore, similar studies
shou Id be carried out also for several other open star
clusters and additional observations in all fields investi-
gated so far, and an extension of the observing programme
to other open clusters and stellar associations are there-
fore highly desirable. One of the principal aims would be to
31

Fig. 2: The tigure at top represents a turther enlargement ot plate
GPO 2557, taken on March 16, 1977, showing the region around
HO 95878 (see tig. 1). The tigure below shows the same section on
plate GPO 2564, taken in the next night on March 17, 1977. Note
the large relative Doppler shitt ot spectral lines ot HO 95878
between the two observations (as compared with the line posi-
tions otsurrounding reterence stars). Measurement yields a ditter-
ence in radial velocity ot 165 km s ' with an expected error ot
a single observation ot ± 7 km s '. This star is one otthe numerous
spectroscopic binaries, discovered by the author on GPO plates.
compare the resu lts for different clusters and eventually to
correlate them with other physical parameters of these
aggregates. In this context a comparison of the kinemati-
cal properties of stellar associations and open star clusters
would be of special interest, because it may provide new
insight into the dynamical evolution of these stellar
systems. Beyond this, the discussion of the velocity
distribution of the field stars and its dependence on
spectral type and distance of the stars may provide
important contributions to our knowledge of the kine-
matics and dynamics of our galaxy. For all kinematical
investigations just mentioned, radial velocities are of
special significance, if they can be combined with reason-
ably accurate proper motions to yield the space velocities
of the stars.
Concerning the investigation of spectroscopic binaries,
additional observations are also desirable: If for example
membership probabilities for a larger number of open
clusters are determined, the interesting question on
possible differences of the binary frequencies in open
clusters, stellar associations and the general star field can
probably be answered. Furthermore, on the basis of
a sufficiently extensive observation material, the correla-
tion between binary frequency and spectral type and
eventually other physical parameters of the stars can be
investigated. Finally, continued observation of all star
fields is desirable as weil for improvement of the orbital
elements of known and newly discovered spectroscopic
binaries as for detection and investigation of long-period
spectroscopic binaries (periods larger than about
200 days). For the latter a good detection probability can
be predicted, since after construction of normal poi nts, the
individual identification of long-period variability of the
radial velocity may be reliable down to radial velocity
amplitudes of only K = 4 km s '.
32
+100 r-~--~--~--~--~----------,
km 5- 1
6V
10
o
-100 L-~_.l...-~_.l...-~_~~_~~_~ _ _.l.-_-----J
o 0.2
Fig. 3: The relative radial velocity curve ot HO 95878 as derived
trom the measurement ot 17 GPO plates of the field around the
open star cluster NGC 3532. The relative radial velocities deter-
mined trom the spectra shown in tigure 2 are identified byarrows.
The zero point is arbitrary, but can be calibrated easily by one
accurately known radial velocity in the tield. The cross marks two
independent measurements.
The application of the objective prism for the determina-
tion of the radial velocities of the stars has opened new
frontiers. The future will show how successful this instru-
ment will be to contribute to the answers of numerous
questions, a few of which have been outlined above.
Since 1977 all observations have been obtained by ESO
night assistant Gorki Roman. I herewith gratefully ac-
knowledge his very careful conduction of my observing
programmes at the GPO.
Assembly of the Coude
Echelle Spectrograph (CES)
0. Enard, ESO Optics Section, Geneva
The coude echelle spectrometer has al ready been de-
scribed in the MessengerNo.11 of December 1977. It is
a very high resolution spectrograph with a resolving power
up to 100,000 fed by either the 3.6 m or the Coude Auxiliary
Telescope (CAT). Henceforth it will be complemented by
the CASPEC which will provide a resolving power of 20 to
60,000 (cf. page 27).
After considerable delay in the delivery of several
components, the assembly of the CES started early in April.
The instrument is being assembled and tested in the ESO
Optics Laboratory in Geneva, which, as a matter of fact, has
become rather congested, as shown by the photographs.
Because no equipment has yet been installed at the
3.6 m coude focus, the whole instrument including the
computer, is assembled in Geneva for a complete test
before it will be shipped to Chile. However, thetemperature
stability of the laboratory is not very good and may
somewhat limit the measured performance level.
Fig. 1.

Figure 1 shows the CAT focal reducer being tested. After

completion of some minor modifications, the focal reducer
will be shipped to Chile where it will be aligned on the
common 3.6 m/CATtelescope axis defined by the "centre"
of mirror 5 of the 3.6 m and the "centre" of mirror 3 of the
CAT.

Fig.3.

Fig.2.

Figure 2 shows the marble supporting the entrance slit

and the associated facilities (calibration, viewer, etc....)
and the pre-disperser. One can see the entrance slit (1), the
collimator unit (2), the prism unit (3) and the exit slit (4), the
double pass system (5) and the scanner detector (6).
Figure 3 shows the grating on its turn-table and Figure 4
the collimators and the multichannel camera. The Digicon
detector will be attached below the camera. Fig.4.
The optical alignment and completion ofthe last mecha-
nical modifications will continue until early July. Complete
integration with electronics and software and initial instru-
ment assessment will take another 2 to 3 months, and the OUTTHERE
official presentation to the Review Team will take place in
October. Installation on La Silla could start in January 1980 Beyond the Milky Way?!
and, therefore, the CES and CAT could become available at How far da bars extend?
the same time, i.e. in April 1980. (ESO Scientific Preprint No. 53-Title)

33
Television Detector System Development at ESO
P. Grane and w. Nees

Science Laboratory at the University of Utrecht in Holland.

The traditional picture of an astronomer with A very similar tube will be flown by NASA in conjunction
a long beard, gazing through his telescope and with the Solar Maximum Mission. Although this tube will
solving the riddles of the Universe, is no longer not be flown on the Space Telescope, a similar larger
true. First of all, the beard would all too easily get version was a very strong competitor tor the Wide Field
Camera Instrument on the Space Telescope. The basic
mixed up with the computer terminal keys and be
detector on the IUE (International Ultraviolet Explorer
adefinite danger near rapidly spinning magnetic Satelli'te) is also a SEC-type tube but of considerably
tapes. And nowadays few astronomers really look different design.
through the telescopes; the important task of Although the system which is being built at ESO is
detecting light from (faint) objects is more accu- primarily aimed at developing the SEC-tube for the
CASPEC application, the electronics have been designed
rately and efficiently done by e/ectronic detec-
in a very general way so that a wide variety of TV-type
tors. Dr. Phi/lip Grane and electronics engineer detectors could be used with the same basic electronics.
Walter Nees at ESO in Geneva are deve/oping For example, silicon vidicons, silicon intensified tubes
such a detector and exp/ain how it works and what (SIT) and many othertubes could be used in awidevarietyof
it can do. operating modes. The electronics are designed to interface
through standard CAMAC binsandseveral special-purpose
CAMAC modules have been constructed. The entire detec-
We describe below the programme of developing a televi- tor system is designed for completely remote control and
sion-type detector system that has been going on at operation.
ESO/Geneva for about 2 years. The major emphasis of
this programme has been to provide a modern detec- How the SEC Works
tor system for the Cassegrain Echelle Spectrograph
(CASPEC), so much of the following discussion We describe below some of the details of television tubes
relates specifically to this particular application. and some of the applications which the ESO device will
Television systems have many attractive features for
optical observations. They allow efficient detection of rable 1: Characteristics 01 SEC-Vidicon Tube
photons. In fact, many television detectors permit the
detection of photons to be recorded with signal levels Tube type: Westinghouse SEC-tube
WX-31958
which are high enough that the resulting data are photon-
noise limited. Another advantage of television detector Photocathode: Tri-alkali with S-20 spectral
systems is that they allow the observer to review his data response. Quantum efficiency
- 18%at4400Älallingto-1%
very soon after they have been obtained. Si nce the data are
at 7500 A.
usually recorded in digital format, the reduction by com-
Focus: Image and beam section
puter techniques is facilitated.
magnetically
Dellection: Magnetically
A SEC 10r the CASPEC
Target storage area: 25 x 25 mm square
For the above and other reasons, it was decided to build Resolution: At MTF 30%
a television detector system. At about the same time, the 1,000 TV-lines/picture height
need arose for a detectorto be used in conjunction with the or 20 LP/mm equivalent
CASPEC, cf. page 27. Apart from photographic plates and Target gain: - 60
conventional image tubes, one television-type detector
Dark current integration: - 2 pAlmin
would fill the requirements of the CASPEC, a 25 mm
magnetically focused "SEC"-Vidicon tube manufactured
by the Westinghouse corporation. The term "SEC" stands
for Secondary Electron Conductivity and refers to the
technique for storing the photon signal internally in the
tube (see below for more details). For the CASPEC
application, the major advantages of the SEC tube are the
tlat image plane, the high resolution and low distortion,
and the capability of operating without significant cooling.
Some technical details of the SEC tube are given in table 1.
A picture of one of these tubes is shown in figure 1.
The particular tube which has been chosen was original-
Iy developed at the Princeton University Observatory
through contracts with NASA for space-borne applications
such as the Space Telescope (ST). The same tube with
a different front window and photocathode is being flown
in a balloon UV-echelle spectrograph by the Space Fig. 1: A SEC tube.

34
find in the Cassegrain Echelle Spectrograph. There are, in WlZslinghouslZ WX 3195B SEC-Vidicon
general, two classes of television camera tubes: The
vidicon plumbicons which store one electric charge per
detected photon, and the SEC- and SIT-vidicon tubes
which provide aprestorage gain mechanism whereby
many (10-500) electric charges are stored per detected
photon.
The basic parts of a SEC-type tube are: the image
Incidcmt
section, the target area, and the electron gun readbeam photons
section (see fig. 2). Incident photons release photoelec-
trons at the photocathode. These photoelectrons are
accelerated towards the target in an electric field and

,- ;
magnetic focus
coil
focused in a magnetic field. They strike the target with
TARGET 5TRUCTURE
about 7,5 Kev of energy and release between 50 and 100
secondary electrons. The target consists mainly of a fluffy

J
layer of potassium chloride (KCI) supported on an alumi- F,.ld mosh
, ,
nium oxide (AI 2 0 3 ) substrate, but separated from it by an , '

~ (l0~)
aluminium signal plate. The secondary electrons are
collected on this signal plate leaving behind in the KCI an \" poroul KCI target layer
electrical image of the distribution of incident photons. By 700}, AI 20 3 supporling layer L 700}, AI signal plat.
scanning a finely focused electron beam ac ross the target,
Fig. 2: Schematic drawing of the main parts of a SEC tube.
it can be recharged and by detecting the recharging
current, it is possible to derive a signal which is propor-
tional to the total number of photons incident at the
corresponding point on the photocathode. Wh at the SEC May Do
Since the recharging currents are often extremely small,
it is of utmost importance that the first stage of the video The data coming from this detector-spectrograph combi-
nation will provide new information at high resolution on
amplification chain have an extremely low noise. The
faint objects. Thus, for example, the chemical composition
system being built at ESO will have an equivalent noise
of galaxies and the stars and nebulae wh ich make up these
level of less than about 5 photoelectrons per resolution
galaxies can be studied to even greater distances. Maybe
element referred to the photocathode.
these data will provide clues to evolutionary processes in
The television tube will be used in the integrating mode
galaxies? Other observations might include the study of
where the read beam is turned off during the time an
absorption-li ne systems in quasars with higher spectral
exposure is being made and the image section is turned off
resolution than previously possible.
during the time the tube is being read out. The tube will
also be run in the "slow-scan" mode in which it takes These applications require extremely sensitive and
roughly 20 seconds to scan and read out a complete 1,000 efficient instruments such as the one described here. Of
course it is too soon to be sure, but we hope that the
element by 1,000 line frame. This operating mode can be
combination of a modern detector, an efficient telescope
compared to the continuous scan 25 frames per second
and spectrograph and fast efficient data-reduction tech-
operation of a commercial-type television camera. Clearly
niques will make this into one of the most productive
the electronics to control an integrating slow scan system
astronomical instruments on La Silla or elsewhere.
will not be very similar to a commercial television camera.
A typical operating sequence would be:
1. PREPARE, the target is erased of any previous residu<;ll
exposure and prepared to a fixed charge
3900 3951
state; 4009
2. EXPOSE, the read beam is off, and the photocathode
high voltage (- 7,5 KV) is on. Photoelec- -
trons are storing up a charge image on the
-
target;
-
3. READOUT, the electron beam scans the target. The
-
measured target recharge signal is, after
-
4767
amplification, sampled and converted into K =31 4844 4932
12 bit binary words at intervals of about 15
microseconds each. (1,000 words per scan -
line and 1,000 scan lines per complete -
frame). For further processing the digital
-
image data 1,000 x 1,000 or 10· words are
stored on computer disc, or magnetic tape. --- 5460
5561
The Cassegrain Echelle Spectrograph will make full use
5666
of the 25 mm square active area of the SEC tube. The
optical system of the CASPEC has been designed to take
advantage of the resolution of the SEC, so that a 1.3 arc
second slit will project into 30 microns on the SEC Fig. 3: Schematic of an echelle spt;ctrum. K is the echelle grating
photocathode, when the R = 17,000 grating is used. order. Wavelengths are shown in Angströms. Intermediate orders
Figure 3 shows a typical vidicon frame (simplified) as it are shown by tick marks. The circumscribed square is the
would appear to the CASPEC user. 25 mm x 25 mm target area of the SEC Vidicon.

35
Collaboration on the Use of the 4 cm and 9 cm McMulian
Electronographic Cameras at the Danish 1.5 m Telescope

K. Gy/denkerne, R. F/orentin Nie/sen and D. McMullan

Two electronographic cameras have now been

operating at the Oanish 1.5 m telescope on La
Silla du ring several months. They assure an
efficient use of this fine telescope and many
exciting photos have already been obtained. Ors.
K. Gyldenkerne and R. Florentin Nielsen of the
Copenhagen University Observatory and Dr. 0.
McMullan of the Royal Greenwich Observatory
explain how these cameras work and inform
about some of the far-reaching observing pro-
grammes that have been initiated.

The commissioning of the Danish 1.5 m telescope at the

European Southern Observatory (ESO) at Cerro La Silla in
Chile has been reported by Andersen, Florentin and
Gyldenkerne (1979), who described the first tests of the
telescope and gave a brief summary of its auxiliary
instrumentation. Additional test periods include various
observational programmes using direct photography,
spectro-photometry, and photometry with these instru-
ments. In particular, the 4 cm and 9 cm electronographic
cameras developed and constructed by D. McMullan and
his collaborators at the Royal Greenwich Observatory
(RGO) (McMullan et al. 1972, 1976; McMullan and Powell
1976,1979) are being used extensively for direct electrono-
graphy. The observations with these cameras on the 1.5 m
telescope are part of a collaboration between the RGO and
the Copenhagen University Observatory (CUO).
A basic requirement for the optical specification of the
Danish 1.5 m telescope was that it should have a Ritchey-
Chretien mirror system with a useable field of a little less
than one degree and thus be complementary to the ESO
1.5 m spectroscopic telescope. Furthermore, it was antici-
pated by the initiators of the telescope project, Professors
A. Reiz and B. Strömgren, that an electronographic camera
would be available with such a large cathode area that it
would cover the uncorrected 80 mm diameter (20') Casse-
grain field.
During the telescope construction period it became
c1ear to the Danish scientists responsible for the auxiliary
equipment that the electronographic cameras developed
at the RGO would be very suitable for the 1.5 m telescope.
After the Austin conference on "Electrography and Astro-
nomical Applications" in 1974 (Chincarini, Griboval and
Smith 1974) it was decided to initiate a collaboration
between the RGO Physics Laboratory and the CUO
Astroelectronics Laboratory on the further development of
the cameras. The principal goal was to provide a Danish
4 cm camera and to use this camera and the first RGO 9 cm
camera on the Danish telescope during its test period. In
the winter of 1974-75 R. Florentin Nielsen spent six months
at the RGO Physics Laboratory and contributed to the
completion of the first 4 cm camera, which was then taken
Fig. 1: The RGO 4 cm electronographic camera (schematically).
to the Wise Observatory in Israel for regular observational
use. Then, in 1976-78, a Danish engineer, Finn Johannes-
sen, participated in the camera projects at the RGO for
about 18 months. Finally, the present authors commenced
the tests of the two cameras on the 1.5 m telescope in the
fall of 1978. The cameras, the observational programmes
for the test period, and some of the results are described
here.
The Electronographic Camera
The performance characteristics of an electronographic
system can be summarized as folIows:
(1) The detective quantum efficiency may approach the
responsive quantum efficiency of the photocathode, be-
cause every photoelectron entering the emulsion leaves
a developable track.
(2) The spectral range is wide, depending on the
characteristics of the photocathode, with quantum effi-
ciencies significantly higher than with photography over
the entire spectral range (3000-9000 A).
(3) There is no reciprocity failure.
(4) The detection process is linear; that is, the density of
the developed image increases linearly with exposure, up
to densities of 4 or 5 in the case of certain nuclear
emulsions.
(5) The dynamic range of nuclear emulsions is high
because of their fine grain and the possibility of exposing
to high density. The fog level is very low.
(6) The spatial resolution can be better than what is
needed in conventional astronomical recording.
The historical development of electronographic came-
ras will not be described in detail here but mention must be
made of the pioneering work of A. Lallemand; however, the

36
operation of his Camera Electronique is rather compli-
cated and time-consuming. Another pioneer, J. D. McGee,
developed the "Spectracon" mica window image tube
which is very easy to use and is commercially available but
has the disadvantage that its photocathode area is limited
to 25 x 15 mm 2 • The cameras developed at RGO are also of
the mica window type but the design permits the incorpo-
ration of larger photocathodes, 44 mm and 93 mm diame-
ter respectively in the two avai lable versions.
The 4 cm camera is shown schematically in figure 1. The
photocathode is normally of the S.20 type which has
a spectral response extending into the near infrared;
ideally it should be formed on the silica glass faceplate but
for the time being aseparate thin silica substrate is used
because of technical difficulties (which howevershould be
overcome soon). The processing of the photocathodes is
carried out in a way which prevents contamination of the
tube interior with alkali metals and results in very low dark
current. The photoelectrons are accelerated to 40 keVand
focussed by parallel electric and magnetic fields onto the
mica window which is 40 mm diameter and 4 ~lm thick. Its
purpose is to isolate the photocathode from the gases
evolved from the nuclear emulsion which would cause an
immediate loss of photosensitivity, while permitting a large
proportion of the accelerated photoelectrons to pass
through. The window is protected from atmospheric
pressure by a vacuum lock through which the electrono-
graphic film (nuclear research emulsion on a Melinex
base) is introduced and pressed into contact with the mica
by an air pressure of - 15 Torr. The resulting mechanical
stresses in the mica are very small and there is no danger of
breaking the window.
The loading and unloading of the film through the
vacuum lock is carried out by an automatic electropneu-
matic system which incorporates a number of safety
interlocks to safeguard the tube against operator error or
faults in the control system. The films commonly used are
IIford L4fine-grain nuclearresearch emulsion and the more
rapid IIford G 5 emulsion. The resolution under optimum
conditions is 60 line pairs per mm corresponding to 10 ~m
or 0.15 arc second at the 1.5 m telescope f/8.5 Cassegrain
focus.
The tube is of demountable construction, so that in the
event of failure, for example of the photocathode, the tube
can be reprocessed easily (and at a moderate cost).
However, because of the stringent leak testing procedures
that are followed, the photocathode should have little loss
in sensitivity over aperiod of years unless a small leak
opens up or the cathode is damaged by exposure to too
high a light level.
Operation of the tube is easy, and it can be used under
the most humid conditions and at observatories at the
highest altitudes. Further details regarding the use of the
tube and its maintenance are given in an "Operational
Manual" and a "Maintenance Manual". The loading of the
film and the dark-room work are done as in normal
photography, and the development and fixing techniques
are similar to those for 111 a-J plates.
In the 9 cm electronographic camera the diameter of the
mica window is 85 mm but a demagnification in the
electron optics makes the useful cathode diameter93 mm.
Since it has not so far been possible to make the large tube
envelope of fused silica (as with the 4 cm tube) Pyrex glass
is used and this causes a lower sensitivity in the ultraviolet.
Otherwise the construction of the larger tube differs only in
minor respects from the small one, and the same control
electronics is used for control and supply of both cameras.
A standard set of optical filters is provid~d and mounted
in two wheels in front of the camera. Large area filters
which can be used with both cameras include the standard
Johnson broad-band U, B, V filters, a red filter, and a "dark
sky blue" (DSB) filter. The latter has a very sharp bandpass
transmission curve 800 A wide centred at 4900 A and is
used in order to reduce the effect of night sky emission
lines. In addition a set of standard Strömgren intermediate-
band fi Iters (u, v, b, y) and another red fi Iter are available for
the 4 cm camera only. There are also spare positions in the
smaller wheel for special filters to be used with the 4 cm
camera.
The Observations
Various electronographic programmes were planned for
the test period; we shall briefly mention these and describe
a few results in more detail.
In the initial test period in November 1978 R. Florentin
Nielsen in collaboration with Karen T. Johansen observed
aseries of first rank EO galaxies in clusters of galaxies,
mainly of the Bautz-Morgan type 111 having similar rich-
ness. This is a pilot programme for the study of the
evolution of elliptical galaxies for selection of an evolution
parameter in cosmological distance scale determinations.
In a search of faint extensions of galaxies (Disney 1976)
R. Florentin Nielsen initiated a galaxy morphology pro-
gramme. The galaxies observed cover a wide range of
types und were selected from the Hubble Atlas. Observa-
tions were made partly with the 4 cm and partly with the
9 cm camera to match the angular size of the galaxies.
V and DSB filters, and in a few cases also Band R, were
used; the DSB filter was particularly useful in obtaining
maximum contrast to the sky background and thus
reaching a very low limiting surface brightness.
In the second moon-free period K. Gyldenkerne in
collaboration with J. G. Bolton (Parkes) and co-workers
made electronographic observations (V and DSB filters) of
selected faint radio sources for a more precise morpholo-
gical classification of optical counterparts than what is
possible in normal photography. In another dark period
P.A. Wehinger (Heidelberg), Susan Wyckoff (Ohio State
University) and K. Gyldenkerne using a special filter
(Schott OG 5 70) searched for underlying structure sur-
rounding low redshift quasars and also studied structure
and orientation of radio galaxies in the optical area with
respect to their radio structure, i.e. double and tripie radio
lobes.
B. Thomsen and S. Frandsen (Aarhus) also studied
"metric" diameters of galaxies belonging to selected
clusters. In addition, they carried out U, B, V, R photometry
(R red filter) of BL-Lac objects with the purpose of
detecting outer areas of galaxies possibly surrounding
these objects. Thomsen and Frandsen are also studying
globular clusters belonging to nearby galaxies (B, V, R) in
order to compare the spatial distribution with the surface
density of the other halo stars, to determine the luminosity
function for globular clusters belonging to different gal-
axies, and, if possible, to derive a correlation between
cluster colours and distance from the galaxy centre.
P. Grosb01 made u, b, y, R observationsofspiral galaxies
in order to compare the observed colou r distributions with
those calculated theoretically by himself and C. Yuan on
the basis of the density-wave theory.
The above observations were all made on dark nights. In
addition, standard sequences were observed in various
fields. Thus K. Gyldenkerne in collaboration with J. G.
Bolton, R. Cannon and A. Savage started observations of
37
sequenees down to 20 mag for the applieation to the
quasar seareh programme deseribed by Bolton and 8av-
age (1978).
The first moon period in Deeember 1978 was taken up
with a programme of narrow-band eleetronography of
hot-spot galaxies. The work is a eollaboration between
D.J. Axon (University of 8ussex), K. Taylor (RGO) and
K. Gyldenkerne.
The first results of this programme were most promising.
The nueleus of NGC 1808 was studied in detail, eleetrono-
graphie images being obtained in the majority of the
prominent emission-lines (Ha, [N 11], [811], [Oll], [0111], Hß,
ete.) together with some intermediate eontinuum bands.
Examples of these data, whieh are being prepared for
publieation (Gyidenkerne, Axon and Taylor 1979) at this
time, are presented in figure 2.
As further indieation of the power of this teleseope-de-
teetor system, we obtained sky-limiting exposures of NGC
1808 in the Johnson V band and as a result diseovered
a beautiful system of filaments emanating from the dise

• •

a b c

(,

• • •

d e f

<"'"9 1
h
Fig. 2: Narrow-band electronographic exposures of the active galaxy NGC 1808, obtained with the 4 cm McMullan camera on the Oanish
1.5 m telescope. Emulsion lIford G 5. Filter halfwidths around 15 Ä. (a) Negative 6547: [Oll), 60 min; (b) Neg. 6546: [Olll}, 60 min; (c) Neg.
6551: continuum near 6400 Ä; 30 min; (d) Neg. 6545: Ha, 60 min; (e) Neg. -6550: Ha, 180 min; (f) Neg. 6548: Ha, 360 min; (g) Neg. 6556:
[N II}, 90 min: (h) Neg. 6544' [S II}, 30 min; (i) Neg. 6549: [S II}, 150 min. Note the difference between the /in es, due to differences in
temperature and pressure in the central regions. The Ha photos show H 11 regions. There are some unavoidable plate faults; the "comet ..
in (h) is one of these.

38
region of the galaxy, very similar to the filaments in
that other famous southern hot-spot galaxy NGG 1097
(Wolstencroft and Zealey 1975).
Summary
The experience with the 4 cm and 9 cm cameras in
combination with the 1.5 m telescope has been extremely
satisfactory. The power of the large camera is illustrated by
figure 3 which should be compared with the 30-minute
electronograph of the same area shown in Messenger
No. 16, p. 1. The IIford emulsions used from the beginning
of the test period have turned out to be of good quality and
with very few defects occurring, so that practically all the
films provided could be used.
With the good seeing experienced in the 1.5 m telescope
dome and the very fine external seeing occurring occa-
sionally, coupled with the high quality of the telescope
optics (Andersen and Niss 1979), the electronographic
resolution mentioned above will permit a very faint limiting
magnitude with this combination of instruments. Limits of
about 26 magnitude for stars and 27 'magnitude per
square arc second for extended objects should be ob-
tained under optimum weather conditions.
The collaboration described has thus been successful. It
should be added that the achievements during the period
in question have also, at least partly, encouraged the
production of similar electronographic equipment for
other non-British telescopes such as the ESO 3.6 m
telescope on La Silla.
Acknowledgements
The authors thank staff members at RGO and at GUO for all
their assistance. The Danish Natural Science Research
Gouncil awarded a special grant for the Danish part of the
project and financed Finn Johannessen's work at RGO.
A NATO Research Grant has provided essential support for
the Danish-British collaboration.
References
Andersen, J., FIorenIin, R. and Gyldenkerne, K. 1979 The Mes-
senger, 16, 1.
Andersen, J. and Niss, B. 1979, The Messenger, 17, p. 14.
Discovery of a New Eclipsing
Dwarf Nova: OY Carinae
The weather at the beginning of the night April 29/30, 1979
was not excellent on La Silla: Most clouds were gone, but
many of them were just waiting near the horizon and
threatened to come back. After twilight I pointed the ESO
1 m telescope to the southern dwarf nova OY Gar, which is
a faint variable star of about 16~. The first photoelectric
measu rement, however, revealed 14~'8, a bit brighter than
normal, and it seemed to brighten up rapidly. Of course,
this is not unusual for a dwarf nova: OY Gar just was
beginning one of its eruptions. lieft the telescope on the
star monitoring it continuously in 3-second time intervals.
Fig. 3: A 90-minute exposure of a region around NGC 2081 in the
LMC, taken with the 9 cm McMullan camera on the Oanish 1.5 m
telescope. Emulsion lIford G 5 and filter "Oark Sky Blue". The
exposure may be compared with the one shown on page 1 of
Messenger No. 16.
BoiIon, J. G. and Savage, A. 1978, lAU Symposium No. 79, 295.
Chincarini, G. L., Griboval, P.J. and Smilh, H.J. (ed.) 1974,
Proceedings, Conlerence on Eleclrography and Aslronomical
Applications, Universily 01 Texas.
Disney, M. J. 1976, Nature, 263, 573.
Gyldenkerne, K., Axon, D. and Taylor, K. 1979, in preparalion.
McMullan, 0., Powell, J. R. and Curlis, N. A., 1972, Adv. EEP, Vol.
33A, p. 37.
McMullan, 0., Powell, J. R. and Curtis, N. A., 1976, Adv. EEP,
Vol. 40 B, p. 627.
McMullan, D. and Powell, J. R. 1976, Proceedings, lAU Colloquium
No. 40,5.
McMullan, D. and Powell, J. R. 1979, Adv. EEP, in press.
Wolstencrofl, R. D. and Zealey, W. J. 1975, Mon. Not. R. Astr. Soc.,
173,91 p.
After five minutes, however, a new surprise showed up: the
intensity began to drop! I checked the diaphragm: the star
was properly centred. I looked at the sky: No visible clouds
in the field. But the signal diminished more and more, and
within a few minutes the star was five times fainter than
before! This did not last long: the intensity began to rise
again, even faster than the decline, and reached its
previous value five minutes later. Gould this have been
caused by the unstable weather conditions? I kept the
telescope on OY Gar and found out that the darkening of
this star repeated periodically every 91 minutes: no cloud
is known to pass so regularly!
There was no doubt that I had discovered a new ecl ipsi ng
dwarf nova with an extremely short period. Figure 1 shows
the first eclipse ever observed: the star was still faint (-15 m )
at this time. Two totality phases appear like "steps" in the
lightcurve. According to recent dwarf nova models, the
first "step" corresponds to the total eclipse of the central
parts of the disk and of the white dwarf, while the second
39
 and deepest "step" coincides with the
coverage of the "hot spot". The
second lightcurve shown in figure 1
corresponds to the next night when
OY Gar had reached the maximum
brightness of its eruption (-12':'4).
Only a partial eclipse is observed be-
cause the eclipsed body (the disk) is
now much more extended than before,
and cannot any more be covered total-
Iy by the faint red secondary star.
Since we see the hot spot radiation
separated from that of the disk at
certain eclipse phases, we can calcu-
late the relative contribution of both
components to the total light, and can
follow up this ratio throughout the
outburst. This will have important con-
sequences for the dwarf nova outbu rst
mechanism which is still not definitely
known.
The discovery of this eclipsing bi na-
ry was not a pu re accident: in January
1979, when I took spectrograms of
several dwarf novae with the Image
Dissector Scanner at the 3.6 m tele-
scope, OY Gar turned out to show
a strong, double Balmer emission with
aseparation of -1,500 km/s of both
emission peaks. This is typical for
cataclysmic binaries with high orbital
i nclination and, thus, justified a search
for eclipses. The rest was good luck:
fair weather conditions, the begin of
ALGUNOS RESUMENES
La calidad dei telescopio
danes de 1.5 m sobrepasa
las expectaciones.
Desde fines de noviembre dei ario pasado
un nuevo instrumento se eneuentra a dis-
posiei6n en La Silla: EI teleseopio danes de
1.5 m.
En una serie de plaeas tomadas durante
una noehe de exeelentes eondieiones at-
mosferieas a prineipios de marzo dei pre-
sente ario, el tamario de las imagenes
variaba de 1 segundo de areo hasta 0.6 se-
gundo de areo. La mejor plaea-eon una
exposiei6n de una hora-muestra bonitas
imagenes eireulares de 0.5 segundo de
areo! Es este un resultado easi inerefble,
que eomprueba la exeelente ealidad deeste
nuevo teleseopio.
ö Crucis es variable!
Durante una reeiente estadia en La Silla el
Dr. Erie W. Eist dei Observatorio Royal en
Ueele, Belgiea, ha deseubierto que una de
las estrellas en la Cruz dei Sur, Ii Crueis, es
variable.
Naturalmente hay muehas otras estrellas
variables, sin embargo. el presente easo es
partieularmente interesante porque la am-
plitud maxima en la eurva luminosa tiene
40
: '.".
'.'
;,':'
.'~ ......... /.",,':'
....'..
",
-.
OY Car
TIME
Fig. 1: Eclipse lightcurves of OY Gar,
observed on April 29/30, 1979 (uppercurve)
and April30/May 1 (Iower curve) with the
ESO 1 m telescope. The uppercurveshows
two totality phases which are indicated by
bars (explanation see text). Each curve
covers a total time interval of about 13 mi-
nutes.
an o'utburst and-Iast but not least-
an excellent cooperation of Holger
Pedersen who was scheduled for half
of the observing nights, but kindly left
me the critical hours for OY Gar.
Nikolaus Vogt
una magnitud de s610 0.006! Esto expliea el
porque hasta ahora la variabilidad de la
estrella no habfa sido deteetada antes,
a pesar de que esta estrella habia sido
observada muehas veees.
EI deseubrimiento es una demostraei6n
de la exeelente ubieaei6n de La Silla y dei
buen rendimiento dei teleseopio de 61 em
de Boehum y su fot6metro, eon el eual se
efeetuaron las observaeiones.
Nuevas series
de diapositivas de ESO
Durante los pr6ximos meses se dispondra
de dos nuevas series de diapositivas.
La primera de estas eonsiste de 20 diapo-
sitivas de 5 x 5 em que muestran las insta-
laeiones en La Silla. Ineluyen edifieios,
teleseopios y vistas dei lugar. Una deserip-
ei6n eompleta en varios idiomas expliea las
diapositivas.
La segunda serie eontiene algunas de las
mejores fotogratras que han sido tomados
eon la eamera dei foeo primario dei tele-
seopio de 3.6 m (eorreetor Gaseoigne). Oe
entre mas de 1.000 fotogratras, se eligieron
20 diapositivas en blaneo y negro (nebulo-
sas, galaxias, ete.).
EI preeio por una serie de diapositivas es
de DM 18.- (0 su equivalente) en Europa,
y US$ 10,- por eorreo terrestre a todos los
demas pafses. 0 US$ 12.50 para su envio
por vfa aerea (pagadero por adelantado).